Printer Friendly

Perceived risk: the link to plan selection and future utilization.

Perceived Risk: The Link to Plan Selection and Future Utilization

Abstract

Enrollment of Medicare beneficiaries in HMOs may reduce the costs of providing health care but has raised concerns over adverse selection and how to adjust reimbursement levels to compensate for this potential. An adjustment based on an individual's perceived health and financial risk may be more accurate than proposed demographic, prior use of health services or current health status adjusters. Results suggest that those in HMOs may be more sensitive to the extreme cost of hospitalization than those with fee-for-service insurance, and therefore were attracted to the comprehensive care offered by the HMO. In addition, HMO enrollees, face fewer barriers to seeking health care, thus may be higher utilizers in future periods.

Introduction

In 1985, out-of-pocket health care costs for the elderly were greater in real dollars than before the 1965 Medicare legislation[7]. This phenomenon did not go unnoticed by the elderly. Realizing they were spending more and more of their budgets on health care, they began to look for insurance mechanisms with which to supplement their Medicare coverage.

At the same time, the Federal government was also experiencing changes in the health care marketplace. The Federal government has seen Medicare expenses increase from $4.7 billion in 1967 to $64.6 billion in 1984[11]. Th is rapid increase in expenditures prompted the Federal government to expand health care delivery mechanisms for Medicare enrollees beyond the traditional fee-for-service (FFS) system through legislation supporting the enrollment of Medicare beneficiaries in health maintenance organizations (HMOs). These trends at the consumer and the payor level have resulted in the increased enrollment of Medicare beneficiaries in HMOs. In 1981, 595,000 Medicare beneficiaries (a little more than 2 percent) were enrolled in capitated systems. In 1985, the number of Medicare beneficiaries enrolled in HMO/CMP and group practice prepayment organizations reached 1,117,000 or 3.7 percent of Medicare enrollees[11].

The initial contracts between the Health Care Financing Administration (HCFA) and HMOs were structured on a cost basis. This type of contract was attractive to HMOs because it limited their risk and thus encouraged them to enroll Medicare beneficiaries. In 1985, HCFA began to move away from the signing of cost contracts with HMOs to the establishment of risk contracts [13]. Under the initial risk contracts the HMOs enrolling Medicare beneficiaries were reimbursed at 94 percent of the AAPCC (Adjusted Average Per Capita Cost). This structure gave HMOs financial incentives to provide services at lower costs (one of their claimed advantages over a fee-for-service practice). At the same time, these contracts would lower the level and increase the certainty of Federal expenditures for health care. Recently, the General Accounting Office (GAO) has recommended lowering the multiplier for the AAPCC to 89 percent[15] thus shrinking the average reimbursement and creating further pressure upon providers (HMOs) to deliver medical services at a lower rate or to exit the Medicare HMO market.

Problem Statement

The move to risk contracts for HMOs creates a number of concerns for the HMOs operating under these terms. The primary concern is whether or not adverse selection will occur. Adverse selection occurs when the high risk/high cost members of a population tend to disproportionately enroll in one type of insurance plan[1]. Recently, HMOs have acted as if adverse selection has occurred in their Medicare population. For example, Av-Med Health Plan Inc. in Florida, Maxicare in California, and Choice Care in Cincinnati have all withdrawn from Medicare risk contracts[14]. Their actions may be in response to adverse selection by enrollees which resulted in the costs of providing services in excess of reimbursement levels. A review of selected empirical studies of adverse selection by HMO enrollees will follow.

The HMO product commonly is presented as being more comprehensive than other supplemental insurance products and thus may attract the higher utilizers. Because of this potential for adverse selection, providers and HCFA have begun to consider prior use of health services and current health status as additional factors in the AAPCC calculation [4, 18, 25]. With these adjustments, it is hoped that the capitated payment will more accurately reflect future utilization. One proposal has been to adjust the payment to reflect the enrollee's prior use of health services. Proponents suggest that an individual's prior use of health services is an accurate predictor of future health services utilization. In addition, for current Medicare beneficiaries, this information is available from claims data. However, while prior use of health services may be an accurate predictor of future utilization when the individual is suffering from chronic health problems, such as diabetes or hypertension, it is a poor predictor for those individuals who will experience acute episodic health care needs.

An alternative to adjustment for prior use of health services has been to base an adjustment upon an individual's health status. Thomas and Lichtenstein[25] examine a variety of health status measures with regard to their ability to accurately predict utilization and costs. These include two perceived health status measures, two functional health status measures and two indirect measures (chronic conditions and prior use of services). Using Medicare enrollees in the state of Michigan, they predict standardized payments for the first six months of 1983 using the demographic factors in the AAPCC (age, sex, institutional status and Medicaid) and their selected health status measures. In all cases, the health status measures dominate the AAPCC factors in terms of statistical significance and size of coefficient and contribution to the explanatory power of the model. Further, indirect measures, prior year's payments and prior year's utilization, explained the greatest proportion of variance in standardized payments. None of the models, however, explained even 7.5 percent of the variance in payments.

As a result of the findings from this and other related research[4, 8, 9, 18], HMOs would receive more reimbursement if their enrollees are in poorer health or used more services in prior years. While this proposal has the advantage of a somewhat closer association to future utilization of health services, it suffers from limitations in regard to measurement. Specifically, debate has ensued over how health status will be measured and who will do the measurement. Clearly HMOs would have incentives to understate their Medicare enrollees' health status, while HCFA has the opposite incentives. Also, as Thomas and Lichtenstein[25] point out, prior use of health services will result in a system similar to cost-based reimbursement with a one-year lag in the estimation of cost. The efficiency incentives from HMO Medicare risk contracts may disappear with this type of reimbursement.

These proposed adjustments are based on a physiological perspective of what determines an individual's need for health services. As an alternative, taking a phychosocial perspective that tries to explain an individual's demand for health care services is proposed. According to the Health Belief Model [19], an individual's perception of risk will influence the decision to seek medical care. In theory, those individuals who perceive a higher level of health risk and/or higher level of financial risk will be more likely to utilize treatment and/or preventive health services.

Therefore, for the purposes of discussion, we propose an adjustment based upon the individual's perceived risk. The objective of this study is to test for differences in perceived health and financial risk between a sample of HMO and non-HMO elderly. To what extent the AAPCC and other demographic factors, prior use of health services, and current health status differentiate these two groups will also be examined. If measures of perceived risk are indeed accurate predictors of the future demand for health services, they may serve as an unbiased adjustments to the AAPCC that will reflect future utilization.

`

Theoretical Background

The pertinent theoretical models in the significant sub-areas are the Health Belief Model and the effects of perceived risk upon decision making.

Health Belief Model

The Health Belief Model proposes the following theoretical conditions and components: (1) an individual's psychological "readiness to take action" relative to a particular health condition is determined by both the person's perceived "susceptability" or vulnerability to the particular condition, and by his or her perceptions of the "severity" of the consequences of contracting the condition; and (2) an individual's evaluation of the advocated health action in terms of its feasibility and efficacy, weighted against his perceptions of psychological and other "barriers" or "costs" of the proposed action. Finally, a "stimulus" either internal or external must occur to trigger the appropriate health behavior[19].

Thus, the Health Belief Model proposes that individuals making health care utilization decisions will evaluate their personal vulnerability, the costs versus the benefits of alternative actions, and will make a health seeking decision in response to some occurrence that motivates the individual to action. A variety of studies have provided empirical documentation of the Health Belief Model. For example, Janz and Becker[17] recently published a review of 46 health belief model studies. They found strong support for the predictive effects of perceived vulnerability on various types of health related behavior. Greater perceived vulnerability led to greater use of preventive medical care, compliance during treatment, and utilization of acute services.

Perceived Risk and Decision Making

In regards to the role of perceived risk upon decision making, Bice has suggested that "risk" vulnerability refers to an individual's expectations regarding his future needs for medical services[6]. HMOs enrolling Medicare beneficiaries under a risk contract (fixed reimbursement on a per member basis) are concerned about whether or not individuals who enroll in the HMO have a higher risk vulnerability than those who do not enroll. From a theoretical standpoint, individuals who are more sensitive to the probability of incurring large, unanticipated medical bills would be attracted to the comprehensive coverage of HMOs[8]. According to economic theories, individuals with a higher degree of perceived risk (loss) will derive more utility from certainty (insurance). Any risk averse individual will be willing to pay a price (the premium) for certainty (insurance) rather than taking chances on alternative incomes having a higher expected value[10]. As the degree of ris k aversion increases and as the probability of an adverse outcome rises, the level of utility derived from certainty (insurance) also increases (although perhaps at a different rate), therefore, the more risk averse an individual is and the greater the perceived risk of illness, the more attractive a comprehensive health plan (HMO) should become.

These theories were applied in a study[3] which examined a group of autoworkers in the Detroit area. The goal was to determine whether differences in demographic and socioeconomic characteristics would explain why only 68 percent of the members of the Blue Cross/Blue Shield health plan were aware they had been given a dual choice option three years ago. Their findings indicated that the older and higher income members of the group were more likely to recall having the option to change plans. The authors speculated that this is due to an increased feeling of vulnerability to risk (illness) and therefore more receptivity to health care coverage options.

More specific applications of these theories to the elderly have also been published. Rundall and Wheeler[22] found that a greater perception of vulnerability to the Swine flu virus (perceived risk) was a strong positive predictor of seeking out recommended preventive medical care (e.g. Swine flu inoculation) for the elderly. Also, Stoller [24] found that elderly people with greater worry about health and concern about the level of health's influence on desired activity led to greater perceived value of physician visits.

The results of these studies as well as the economic theory outlined above support the argument that individuals with a higher level of perceived risk are more likely to be aware of alternative insurance options and to seek out comprehensive health care coverage (HMOs). This supports our suggestion that HMOs will experience adverse selection because those individuals who have greater risk will be more aware of the HMO option and be most likely to demand supplemental health insurance.

This leads to the following hypothesis: an individual's decision to enroll in an HMO or a fee-for-service plan will be influenced by that individual's perceived health risk ("personal vulnerability") as well as by his or her perceived financial risk. Therefore, those individuals with higher levels of perceived risk (health and/or financial) will be more likely to enroll in HMOs.

Empirical Research

Since these hypotheses suggest that HMOs will experience adverse selection, a brief review of some of the research that has been conducted to examine whether or not HMOs experience adverse selection is in order. A study by Berki and other[5] examined a random sample of employees from a Michigan firm offering four choices of health plans (three HMO plans and one Blue Cross/Blue Shield plan). They examined the subjects for differences in (1) need-related variables (health status, health concern, ambulatory utilization/costs, and hospital utilization/costs); (2) patient-physician relationships; (3) correlates of economic risk (per capita income and family income); and (4) sociodemographic characteristics (age, sex, average age of family members and family size). Berki's need-related variables (perceived health risk), exhibited no significant differences between the HMO and the BC/BS subscribers. However, the correlates of economic risk showed significant differences across the groups. The HMO enrollees had higher levels of economic risk (perceived financial risk). These results exemplify the importance of distinguishing between health and financial risk.

A study by Jackson-Beeck and Kleinman[16] examined claims data from 11 employers in the Minneapolis-St. Paul area. The employees had all been given a dual choice option within the last three years. In the study, employees were classified as stayers (remaining in the fee-for-service setting) or joiners (enrolling in the HMO). Their results indicated that the joiners were younger and had lower levels of prior utilization of professional (physician) and institutional (hospital) services. This indicates that for this population, the HMO received enrollees who, if prior utilization is a true predictor of future utilization, will utilize fewer services and thus be lower cost enrollees than those in the fee-for-service setting, the opposite of adverse selection.

A study by Merrill and others[20] examined state employees in Tallahassee, Florida and Salt Lake City, Utah who were given the option of enrolling in an HMO (joiners) or remaining in a fee-for-service plan (non-joiners). The employees were evaluated in terms of their insurance claims history, health status, demographic and attitudinal factors. The results indicated that in the Tallahassee area, the joiners were younger, healthier, and had lower claims expenses in the year preceding enrollment. These findings point to the opposite of adverse selection. However, in Salt Lake City, there were no significant differences between joiners and non-joiners, suggesting that adverse selection did not occur and that adverse selection may indeed be site specific.

Finally, a study by Grazier and others[12] examined Washington state employees who were given three choices: (1) remain in a fee-for-service plan, (2) remain in a closed panel group practice HMO plan, or (3) switch to a primary care network HMO plan. The analysis of factors affecting the choice of health plan focused on need for health services, access, prior utilization, and financial capability. The results indicated that those members of the fee-for-service plan who were female, had younger families, and had lower per capita income were more likely to switch to the network plan. The authors suggest that these three factors may indicate that people selecting the network plan had higher health risk as well as greater financial vulnerability. However, when examining the closed panel members, there was no evidence of differences between stayers and switchers. Thus, there is mixed support for the hypothesis that health risk influences health plan selection and will lead to adverse selection.

Study Model

The findings reported here are part of a larger study funded by the National Institute on Aging. The parent study, "A Panel Study of Elderly Health Beliefs and Behavior," is a three-year panel study utilizing four stages of data collection. The data are being collected primarily via personal interviews and archival analysis.

The study population is composed of three samples drawn from the same geographic area. The first sample contains elderly enrollees in a comprehensive pre-paid health and social services plan (Eldercare) offered by a local HMO. The second sample is comprised of enrollees in the HMO who are not in the Eldercare program. The third sample is composed of local residents who see physicians on a fee-for-service basis.

The larger study is designed to evaluate the impacts of enrollment in the Eldercare program upon (1) health services utilization, (2) health beliefs and attitudes, (3) health awareness, (4) preventive health behavior, and (5) functional health status. The study model identifies individual attributes and social/situational factors as potential moderating variables.

The goal of the study reported here was to examine how specific factors are related to the decision by Medicare beneficiaries to enroll in an HMO or to remain in a fee-for-service setting. Therefore, the samples of the study population were collapsed into two groups. The first group (n = 257) is comprised of Medicare beneficiaries enrolled in the HMO (regardless of participation in the Eldercare program) and the second group (n = 145) is composed of local residents who see physicians on a fee-for-service basis. Specifically, the hypotheses tested in this project were: (1) that the level of perceived health risk will differ between HMO and non-HMO enrollees, and (2) that the level of perceived financial risk will differ between HMO and non-HMO enrollees.

Methods

Measures

Five categories of variables were used in the analyses: demographic, prior use of health services, current health status, perceived health risk, and perceived financial risk. Table 1 presents descriptive statistics for all variables and scales. It also contains tests of the reliability of the scale scores.

The demographic variables of interest in this study were age, sex, occupational level and social network. Age and sex were included because they represent the primary adjustors of the AAPCC. Occupational level was a variable based on responses to questions about the respondent's and spouse's former primary occupation(s). Although responses were grouped into one of four levels: unskilled, blue collar, white collar, and professional, for present purposes a two-level variable is defined as: professional and all others. Occupational level is a proxy for income. Social network (a measure of the respondent's amount and type of social contacts) was included because of evidence[23] that other family members and friends may indeed influence the decision-making process and health care behaviors of the elderly.

Prior use of health services consists of five variables: preventive dental visit(s) in the last six months (yes/no), preventive physician visit(s) in the last six months (yes/no), hospital stay(s) in the last six months (yes/no) number of physician treatment visits in the last six months, and emergency room visit(s) in the last six months (yes/no).

Current health status includes the physical activities of daily living scale (PADL) and the instrumental activities of daily living scale (IADL)[21]. These potential moderating variables were of interest because, as was discussed previously, there have been proposals to use measures of health status to adjust the reimbursement rate for HMOs' Medicare risk contracts.

The remaining variables are the measures of perceived risk. Perceived health risk was used as an indicator of an individual's projected need for health care services. As the individual's level of perceived health risk increases, we would expect him to utilize more health services (preventive and/or treatment). Perceived health risk was measured using a list of sixteen specific health problems. The individuals were asked if they thought they would experience the health problem in the next year (short term health risk) or if they thought they would experience the health problem in their lifetime (long term health risk). A positive response received a score of + 1, a neutral response received a score of 0, and a negative response received a score of - 1. The responses to specific health problems were then summed to form an individual's short-term and long-term health risk scores. Larger values indicated greater perceived health risk.

Perceived financial risk consisted of a respondent's perceived completeness of his or her insurance and their estimates of: the average length of a hospital stay for a person over age 65, the percent of Medicare beneficiaries who have a length of stay in excess of 60 days, and the average hospital per diem. These were used as indicators of the individual's projected demand for comprehensive health care coverage which limits out-of-pocket expenditures. As an individual's estimate of completeness of insurance falls and estimates of the length of stay, likelihood of long stays and the costs of hospitalization rise, one would expect HMO enrollment to become more attractive as a means to reduce the "barriers" or "costs" of providing health services in response to health risks.

Analysis Plan

In the first step of data analysis, statistical tests for differences between the two groups were performed for all potential moderating variables. Scale scores were computed for the social network, physical activities of daily living, instrumental activities of daily living, short-term health risk and long-term health risk measures. Each of these scales was reliable (Cronbach's Alpha [is greater than] .80) with the exception of social network.

The second data analysis step was to test for differences in moderating variables between the FFS and HMO group. We compared social network, physical activities of daily living, and instrumental activities of daily living scales as well as the mean values for age, number of physician treatment visits, number of emergency room visits and current health status between the two groups (FFS vs. HMO) using the t-test procedure. For sex, occupation, preventive dental visits, preventive physician visits, and hospitalization in the last six months, the chi-square procedure was used to evaluate differences between the two groups. A distribution analysis of the continuous variables was performed and a high frequency of non-normality was found. Although this finding suggests a transformation of variables or a non-parametric statistical test, results of these efforts do not differ from the untransformed t-tests and will not be presented.

The third step of analysis tested for differences in the outcome variables, perceived health and financial risk, between the HMO and FFS groups. The mean scale score for perceived short-term and long-term health risk was compared between the two groups (FFS vs HMO) using the t-test procedure. The four elements of perceived financial risk were evaluated independently. The first factor was selected to evaluate the respondent's perception of the adequacy of his or her insurance since a negative response to this item may indicate a barrier of health-seeking behavior. This item was compared between groups using the chi-square procedure. The remaining factors (estimates of utilization and cost) were compared between the two groups using the t-test procedure to see if there were statistically significant differences indicating different financial risk levels. Stepwise regression analysis was then used to test for differences in both health and financial risk factors between the two samples.

Finally, the individual components of perceived health risk were evaluated for differences between the two groups using the chi-square procedure in an attempt to identify specific differences that may be significant on an individual health risk item level but not detectable using a scalar measure. This situation could potentially have administrative consequences.

Results

Table 2 shows the test results for all variables that were thought to be potential moderators of the main effects of perceived risk upon future utilization. Using a significance value of .05, differences between the two groups were observed for occupational level, hospital stay, and current health status (IADL and PADL). These results indicate that if the subject's age, sex, social network and most measures of prior use of health services are used as predictors of future use of health services, one would predict equivalent utilization for the HMO and FFS samples. Therefore, no adjustment to the capitated reimbursement rate, based on these often proposed variables to compensate for selection bias, would be required for this sample. The differences observed indicate that those in the HMO sample had higher current health status, used the hospital less often, and were less likely to have been professional or technical workers, the latter indicating that those in the HMO sample may have had lower average income.

Table 2 also compares the main effect variables, perceived health risk and perceived financial risk, across the two samples. For the measures of perceived health risk we can see that the differences between the fee-for-service group and the HMO group are not significant in either the short-term of the long-term. This suggests that the future utilization of health services will likely not differ between the fee-for-service group and the HMO group.

However, the analysis of results for the measures of perceived financial risk are not as easily interpreted. Statistically significant differences between the two groups occur for the perceived adequacy of insurance and the estimates of the probability of a length of stay in excess of 60 days. Approximately two-thirds (65.8 percent) of the HMO respondents perceived their insurance to completely protect them from out-of-pocket expenses in excess of $2,000 for a prolonged period of illness requiring hospitalization and follow-up care as opposed to 23.3 percent of the fee-for-service respondents. This is not surprising given that HMO enrollees do in fact have greater coverage.

This result indicates that the HMO group faces fewer barriers to the utilization of health services in terms of the relative costs/benefits of care. Therefore, despite equal perceptions of health risk, the HMO members may indeed

use more health services in the future than those in a fee-for-service setting.

Responses concerning the elderly's probability of a length of stay in excess of 60 days suggest that the HMO group perceives this to be more likely for the elderly in the general population than does the FFS group. This suggests that those elderly in HMOs feel there is a greater likelihood of the average Medicare beneficiary experiencing an extended length of stay and therefore this group may be more attracted to comprehensive health care coverage which limits out-of-pocket expenditures (e.g., HMOs). The HMO group is more sensitive to the extreme length of stay and costs of hospitalization, thus supporting the hypothesis that HMO enrollment will be more attractive to individuals with higher levels of perceived financial risk. However, when asked to estimate the average length of stay and average per diem cost of hospitalization there were no differences between the two groups.

Table 3 presents a stepwise multiple regression involving health and financial risk as dependent variables. As is`evident, type of health plan did not enter any of the equations. Therefore, when controlling for other significant variables, there were no differences in any of the risk measures between these two samples. Also, many of the moderating variables which differentiated these samples in the univariate tests failed to be significant explanatory factors for risk. Overall, no more than 6 percent of the variance in the dependent variable is explained by significant factors. All but one of the models is significant (p = .05), however. The pattern of significant variables entering these equations is not remarkable except that no demographic, prior use of service, or current health status measures explained the level of long-term health risk.

Of final interest was how individual components of perceived health risk may differ between the two groups. Table 4 shows that the only health problem for which there was a statistically significant difference between the two groups regarded the onset of anemia in the respondent's lifetime. In this situation, 6.9 percent of the FFS respondents indicated that they expected to experience anemia in their lifetime as opposed to only 2.7 percent of the HMO respondents. This result suggests that the HMO may experience a lower than average increase in office visits for blood tests etc. in the long run if indeed these perceptions prove to be accurate.

In summary, the findings indicate that the sample of the Medicare population enrolled in the HMO versus Medicare beneficiaries in the same area who see physicians on a fee-for-service basis is not different in terms of perceived health or financial risk. Examining the perceived adequacy of insurance coverage leads to a different conclusion and indicates that those in the HMO sample face fewer barriers to care and therefore may be higher utilizers in future periods. The remaining components of financial risk indicate that the members of the HMO sample have equivalent estimates of the financial costs and expected utilization. Other factors must explain why these individuals chose to enroll in an HMO (thus limiting their exposure to financial losses).

Implications

These results indicate that measures of perceived financial risk point to different predictions of future utilization than do indicators based upon prior utilization of health services or current health status. The measures of perceived health risk point to equivalent utilization while the measures of perceived financial risk indicate that the HMO group faces fewer barriers to care (out-of-pocket expenses) and consequently may indeed utilize more health services.

An HMO, therefore, may want to provide educational material, health behavior counselling and an alternative to the utilization of scarce and expensive office visits and physician time. This type of education program may allow the HMO to operate profitably under a Medicare risk contract with a population that apparently faces few financial barriers to care.

Full evaluation of this study's hypotheses will necessitate continuing to follow the study panel to see whether these predictions of future utilization prove accurate and evaluating the accuracy of prior use of health services and current health status as predictors of future utilization. In addition, this data facilities evaluating whether or not this HMO is experiencing adverse selection. Therefore, it constitutes the necessary baseline for meaningful evaluation of proposed adjusters to the AAPCC, an important concern in times of shrinking reimbursement. [Figure 1 Omitted] [Tabular Data 1 to 4 Omitted]

References [1]Anderson, G., and J. Knickman, 1984. Adverse selection under a voucher system: Grouping Medicare recipients by level of expenditure. Inquiry 21 (Summer): 133-145. [2]Asher, C. 1984. The impact of social support networks on adult health. Medical Care 22 (April): 349-359. [3]Bashur, R., and C. Metzner. 1970. Vulnerability to risk and awareness of dual choice of health insurance plan. Health Services Research 5 (Summer): 106-113. [4]Beebe, J., J. Lubitz, and P. Eggers. 1985. Using prior utilization to determine payments for Medicare enrollees in health maintenance organizations. Health Care Financing Review 6 (Spring): 27-38. [5]Berki, S., M. Ashcraft, R. Penchansky, and R. Fortus. 1977. Enrollment choice in a multi-HMO setting: The roles of health risk, financial vulnerability, and access to care. Medical Care 15 (February): 95-114. [6]Bice, T. 1975. Risk vulnerability and enrollment in a prepaid group practice. Medical Care 13 (August): 698-703. [7]Collins, J., and J. McDonald. 1984. Health care for Medicare beneficiaries: The HMO option. Nursing Economics 4 (January/February): 259-265. [8]Eggers, P. 1980. Risk differential between Medicare beneficiaries enrolled and not enrolled in an HMO. Health Care Financing Review 1 (March): 91-99. [9]Eggers, P. and R. Prihodi. 1982. Pre-enrollment reimbursement patterns of Medicare beneficiaries enrolled in "at-risk". HMOs. Health Care Financing Review 4 (September): 55-73. [10]Friedman, M., and L. Savage. 1948. The utility analysis of choices involving risk. The Journal of Political Economy 56 (August): 279-304. [11]Gornick, M., J. Greenberg, P. Eggers, and A. Dobson. 1985. Twenty years of Medicare and Medicaid: Covered populations, use of benefits, and program expenditures. Health Care Financing Review 6 (Suppl.): 13-59. [12]Grazier, K., W. Richardson, D. Martin, and P. Diehr. 1986. Factors affecting choice of health care plans. Health Services Research 20 (February): 659-682. [13]Grimaldi, P. 1985. HMOs and Medicare. Nursing Management 16 (April): 16-20. [14]Group Health Association of America, Inc. 1986a. New AAPCC rates: HMO managers assess impact. GHAA News 27 (January): 1,2,4. [15]Group Health Association of America, Inc. 1986b. GAO recommends 5% cut for AAPCC. GHAA News 27 (August): 1,2,4,7. [16]Jackson-Beeck, M., and J. Kleinman. 1983. Evidence for self selection among health maintenance organization enrollees. Journal of the American Medical Association 250 (November 25): 2826-2829. [17]Janz, N. and M. Becker. 1984. The health belief model: A decade later. Health Education Quarterly 11 (Spring): 1-47. [18]Lubitz, J., J. Beebe, and G. Riley. 1985. Improving the Medicare HMO payment formulae to deal with biased selection. In Scheffler, R. and L. Rossiter (eds.), Advances in Health Economics and Health Services Research Vol. 6 Greenwich, Connecticut: JAI Press Inc. [19]Maiman, L. and M. Becker. 1974. The Health Belief Model: Origins and correlates in psychological theory. In Becker, M. (ed.), The Health Belief Model and Personal Health Behavior. Thorofare, NJ: Charles B. Slack. [20]Merrill, J., C. Jackson, and J. Reuter. 1985. Factors that affect the HMO enrollment decision: The tale of two cities. Inquiry 32 (Winter): 388-395. [21]Pfeiffer, E., T. Johnson, and R. Chiofolo. 1981. Functional assessment of elderly subjects in four service settings. Journal of the American Geriatrics Society 29 (October): 433-437. [22]Rundall, T. and J. Wheeler. 1979. Factors associated with utilization of the swine flu vaccination program among senior citizens in Tomkins County. Medical Care 17 (February): 191-200. [23]Shanas, E., and M. Sussman. 1977. Family bureaucracy and the elderly. Durham, NC: Duke University Press. [24]Stoller, E. 1982. Patterns of physician utilization by the elderly: A multivariate analysis. Medical Care 20 (November): 1080-1089. [25]Thomas, W., and R. Lichtenstein 1986. Including health status in Medicare's adjusted average per capita cost capitation formula. Medical Care 24 (March): 259-275.

Rick K. Homan is a Ph.D. candidate in the School of Public Health of the University of Michigan. Gerald L. Glandon is Associate Professor at Rush-Presbyterian-St. Luke's Medical Center. Michael A Counte is Associate Professor in the Center for Health Management Studies of Rush University.
COPYRIGHT 1989 American Risk and Insurance Association, Inc.
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 1989 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Title Annotation:study of health care costs as related to enrollment in HMOs
Author:Homan, Rick K.; Glandon, Gerald L.; Counte, Michael A.
Publication:Journal of Risk and Insurance
Date:Mar 1, 1989
Words:5658
Previous Article:Stability in the present value assessment of lost earnings.
Next Article:Whole life cost comparisons based upon the year of required protection.
Topics:

Terms of use | Copyright © 2016 Farlex, Inc. | Feedback | For webmasters