Printer Friendly

The effect of the Illness Episode Approach on Medicare beneficiaries' health insurance decisions.

As health policymakers make market-oriented reforms, consumers face an increasingly large and varied range of health care and health insurance options (Mechanic 1989), many of them intended to strengthen the purchasers' incentives to be more cost conscious (Enthoven 1986). According to classic economic theory, a key ingredient for effective operation of market incentives is a well- informed consumer. Currently, however, most consumers are far from well informed about health care and health insurance (Varner and Christy 1986; Hibbard and Weeks 1987).

One major decision made by consumers is the level and kind of health insurance coverage to purchase (Marquis, Kanouse, and Brodsley 1985). The Health Insurance Decision Project (HIDP), conducted at the UCLA School of Public Health in 1986 and 1987, developed a new method called the Illness Episode Approach (Sofaer, Davidson, Goodman, et al. 1990) for providing consumers with information the comparative benefits and costs of available health care coverage options. The method uses common illnesses as the basis for comparing out-of-pocket expenses likely to be experienced by individuals with different kinds of health care coverage.

This article presents the results of a quasi-experimental test of the Illness Episode Approach (IEA), as well as results of a less rigorous test of a well-planned insurance education intervention. In the HIDP, the IEA was applied to health insurance decisions facing Medicare beneficiaries. This group needs to understand not only Medicare itself, but also supplemental coverage sold to individual beneficiaries or available through employers as a retiree health benefit, as well as Medicare "risk-contract" HMOs.

REVIEW OF PREVIOUS RESEARCH

Several studies have shown that beneficiaries have limited knowledge of their Medicare benefits (Lambert 1980; Cafferata 1984; LaTour, Friedman, and Hughes 1986; McCall, Rice, and Sangl 1986). Beneficiaries also lack understanding of their Medicare supplements, often overestimating the protection these supplements provide (LaTour, Friedman, and Hughes 1986). Recent studies of enrollment in Medicare HMOs have not directly assessed beneficiaries' understanding of HMO coverage (Freidlob and Hadley 1985; Farrell 1986; Ward 1987), so there is little information about consumer knowledge of this option.

The influence of consumers' level of information has generally been neglected in previous research on health care coverage decisions. In their 1980 review of the literature on HMO enrollment, Berki and Ashcraft developed a model of consumer decision making that has influenced the design of many studies. They surmised that consumers base their selection of insurance features on their perception of their own health risk, and on the likelihood and effect of health-related financial loss resulting from illness. The model implies that consumers select delivery systems according to their preferences regarding various features of health care. The model does not explicitly include the role of information in coverage decisions, but the authors comment that perceived attributes of plans, rather than their objective characteristics, are the basis of consumer decisions.

Empirical studies investigating factors affecting choice of health plan (Garfinkel, Bonito, and McLeroy 1987; Grazier et. al. 1986; Merrill, Jackson, and Reuter 1985) have not addressed consumers' knowledge or their sources of information. However, a recent study examined the effect of the method of presenting health plan information to Medicaid beneficiaries on their enrollment in HMOs (Andrews et al. 1989); it found that each of five methods was most effective with a different type of beneficiary. Comprehension was significantly related to enrollment decisions; higher comprehension was related to a choice of the HMO rather than fee- for-service Medicaid coverage.

There has also been limited research on efforts to provide health insurance information to Medicare beneficiaries. The characteristics of such efforts vary widely. According to Schauffler (1980), an ideal insurance education effort would employ media outreach to alert the beneficiary to the need for information and to available educational services, and it would then educate through interactive personal contact. Davidson's (1988) review of current efforts found that while some programs follow Schauffler's prescriptions, using the media for outreach and interactive personal contact for education,(1) most programs focus on distributing printed materials and answering direct questions, and do not use personal contact to encourage the application of information. Davidson concludes that many beneficiaries have no access to objective insurance information programs, some have access only to minimal written materials, and a few have access to interactive educational programs. Furthermore, he notes that except for the HIDP, none of the programs has been rigorously evaluated to determine the extent of its effect on beneficiary knowledge or decision making.

THE ILLNESS EPISODE APPROACH

The Illness Episode Approach builds on the idea that people have an easier time making comparisons and choices when they can see the implications of different options in specific situations they might experience. The IEA integrates two theoretical perspectives: social learning theory (Bandura 1977, 1986) and the information processing theory of consumer choice (Bettman 1979). Social learning theory holds that people can learn either by direct experience or direct observation. To learn this way, individuals must be able to anticipate the probable consequences of different decisions; thus, information on such consequences will enhance learning. According to information processing theory, consumer decisions are influenced by the extent to which product information can be acquired and evaluated. Consumers have limited information processing capability and often limited knowledge about alternative choices. Simplifying the presentation of information, helping consumers identify key choice criteria, and making choice criteria more salient and available through the provision of information all become useful strategies to support the consumer's ability to process information (Bettman and Sujan 1987).

The characteristics of health insurance options are complex. Medicare beneficiaries face a large number and wide range of choices. Even the best information available about these choices provides abstract, general information on a wide range of topics, often using technical terms. For example, a typical description of a Medicare supplement will specify whether the Part A and B deductibles are covered, the number of additional hospital and skilled nursing facility days provided, the maximum coverage for any prescription drugs, and ways in which preexisting condition limitations are handled. It is difficult for beneficiaries who do not understand the distinction between Part A and Part B, or exactly how a deductible works, or how a skilled nursing facility might differ from a "nursing home" either to understand these policy features or to anticipate how these features would affect them if and when they were to become ill.

The IEA was designed to provide comprehensible information about the out-of-pocket expenses individuals would likely face for several common illnesses under typical available health care coverage options. While the information developed emphasizes financial consequences, it also identifies other key features of coverage options and provides information on these features. For example, choice of provider is identified as a key feature, so that consumers can be aware of the differences in provider choice between fee-for-service Medicare and risk-contract Medicare HMOs. In sum, the information provided helps consumers anticipate the consequences of choices, presents a simplified set of choice criteria, and provides information on those criteria.

In applying the Illness Episode Approach to the circumstances facing Medicare beneficiaries in Los Angeles County in 1986 and 1987, we selected 13 illnesses common among the elderly that varied in severity and in the mix of services required. We also chose conditions about which relatively high consensus had been reached regarding appropriate diagnosis and treatment protocols.(2) For each illness, we specified diagnosis by ICD-9-CM code and severity level using the SysteMetrics Disease Staging classification system (SysteMetrics 1983a, 1983b). We developed a detailed listing of all of the services that would typically be provided for a patient of a given age and gender.(3) A panel of clinicians operating in different practice settings reviewed these profiles, and agreed with the fundamental course of care presented. Based on their comments, minor changes such as the duration of a visit, the appropriate physician specialty for a visit, or the use of a particular diagnostic test or medication were made. We identified and aggregated the costs for all services and products in each profile, using data on Medicare approved charges, service-specific fee screens for usual and customary charges applied by private insurers, and surveys of local providers. This allowed us to calculate typical charges for each illness in Los Angeles in 1986. By determining the services and costs Medicare would cover for each illness, we could then calculate the out-of-pocket costs beneficiaries would experience if they had Medicare Part A and B coverage only.

After developing a typology of Medigap policies, we selected five that were heavily marketed in the area and whose costs and benefits varied widely. We applied the coverage characteristics of each policy to determine the remaining out-of-pocket costs for each illness under each policy option. We also selected two Medicare HMOs, one with and one without outpatient prescription coverage, and applied their service coverage and copayment features to determine out-of-pocket costs. We used the information generated by these analyses to develop the materials used in an educational workshop, as described in the next section.(4) Table 1 presents the information generated through application of the IEA to one illness, a moderate arthritis, and presented to intervention workshop participants.
Table 1: Sample Out-of-Pocket Cost Information Generated
through Application of the Illness Episode Approach; a Case of
Moderate Arthritis, Los Angeles, 1986 (Typical Total Charges for
One Year: $1,458; Total Medicare Reimbursement: $388)
 Out-of-Pocket Costs
 Without With
Coverage Option Provider Assignment
Medicare Only $1,070 $795
Supplement One(*) 1,010 725
Supplement Two 1,010 725
Supplement Three 660 375
Supplement Four 955 670
Supplement Five 425 425
Medicare HMO One NA[dagger] 631
Medicare HMO Two NA[dagger] 89
(*) Specific policies were identified in participant materials.
Many of these are no longer offered or have been revised.
([dagger]) By definition, all HMO providers accept assignment.


METHODOLOGY

HYPOTHESES AND STUDY DESIGN

We hypothesized that participants in workshops using the IEA, when compared to participants in workshops providing more traditional information, would

1. show greater increases in their knowledge of Medicare and their own health insurance coverage;

2. be better able to make coverage decisions;

3. be more likely to eliminate duplicative coverage;

4. spend less on premiums; and

5. show greater improvements in satisfaction with their health care coverage.

A quasi-experimental design, using randomized assignment of self- selected participants, was used to test the effects of the IEA on these hypothesized changes.

Since the educational workshops attended by controls were identical to the Illness Episode workshops in all details except the educational materials used, the study also permitted us to assess, using a less rigorous, nonexperimental pre- and posttest design, the effect of an interactive educational intervention that met the ideal articulated by Schauffler (1980). We hypothesized that the entire group of workshop participants, intervention and control, would show significant increases in knowledge of Medicare and their own insurance, when measured before and after the workshop.

Using a variety of marketing and outreach techniques,(5) we recruited, for a free, three-hour workshop, Medicare beneficiaries who were trying to make a decision about their health care coverage. Beneficiaries who called to enroll in the workshops were randomly assigned either to a workshop using educational materials based on the IEA analysis or to a control workshop.(6) The control intervention used the best available traditional information on health care coverage available to Medicare beneficiaries: accurate and comprehensive tables comparing Medigap supplements and Medicare HMOs available in Southern California, published by a newsletter called Senior World (Klowden 1986). The Klowden tables provided comparisons of such general features as premiums, coverage of Part A deductible and additional hospital and SNF days; coverage of Part B deductible and Part B copayments; and prescription drug coverage.

Participants were not aware that they were in an "intervention" or a "control" group. Except for the educational materials used, all features of the educational intervention were consistent across the two kinds of workshops. Thus, workshops were the same length; the same community facilities were used; each facilitator handled an equal number of intervention and control workshops; and facilitators covered the same topics, specified in detailed protocols, in the same order.

DATA COLLECTION

Workshop participants completed three questionnaires over the course of the research. A preintervention survey (wave 1) was administered to all participants during the first 30 minutes of the workshop; two postintervention surveys were administered by mail three months (wave 2) and nine months (wave 3) after the workshop.

Demographic information (e.g., age, gender, ethnicity, education, income, occupation, and marital status) was collected at wave 1. Health insurance coverage information was collected at all three waves. Information on knowledge of Medicare and other insurance coverage was collected at waves 1 and 2. Information on health status, recent use of health services, and preferences for various characteristics of health plans was collected at waves I and 3. Information about satisfaction with health care coverage was collected at waves I and 3.

A total of 626 individuals participated in the workshops. Of this group, 561 were age 64 or older at the time of the workshop.(7) Our analysis focuses on this group since they were, or were about to be, eligible for Medicare. Their response rate to the wave 2 survey was 78.3 percent and to the wave 3 survey was 80 percent. Incentive payments of two dollars were included in each mailing. Cover letters were signed by each participant's workshop facilitator to reinforce his or her engagement with the project. Multiple telephone follow-ups were made with nonrespondents.

DEPENDENT VARIABLES AND MEASURES

Knowledge of Medicare and Own Insurance. Medicare knowledge was assessed through five true-false questions related respectively to coverage of hospital stays, eyeglasses, outpatient prescription medications, custodial help in the home, and the definition of physicians' acceptance of assignment. All workshop participants answered these questions. Knowledge of private supplements was assessed through four questions about coverage; benefits questioned included the Part A deductible, the Part B deductible, the copayment for Part B services, and the copayment for extended stays in skilled nursing facilities (SNFs). Only participants with Medicare supplements answered these questions. Similarly, knowledge about HMOs was assessed by asking HMO enrollees three coverage questions: whether they could use any physician; whether the plan covered drugs; and whether the enrollee had to pay the Part A deductible.

To determine the correctness of answers to questions about coverage aside from Medicare, we identified the respondents' private supplement or HMO plan from information they provided, gathered information from the carrier about that policy or plan, and identified the correct responses to our questions. The correct information was then compared to the answers given by the respondent.

For each individual on whom we had sufficient data, we also developed an overall score with respect to his or her knowledge of coverage beyond Medicare. This score was based on answers to questions on private supplements for those with supplements, questions about HMOs for enrollees, and questions about both kinds of coverage for the small number of individuals who both had a supplement and were enrolled in an HMO. Scores were standardized so that zero would represent no correct answers and 100 would represent all correct answers.

Coverage Decisions. To assess respondents' ability to make a coverage decision, we asked them at wave 3 to indicate the status of their decision making. They could mark as many responses as applied: purchased new/different supplement; joined new/different HMO; stopped purchasing supplement; disenrolled from HMO; decided to make no change; or still thinking about the decision. At each wave, respondents were asked to identify their coverage by giving the name of the carrier and the name or number of the policy, or by giving the name of their HMO. They were asked to list all of their policies. To compare the kind of coverage reported by participants, we integrated several data elements to construct a four-level measure of coverage: Medicare only, Medicare with a private supplement, a Medicare HMO, and both a Medicare HMO and a private supplement. In this analysis, we exclude those with Medi-Cal, with no insurance, and those whose only form of insurance was a long- term care, "dread disease," or hospital indemnity policy.

These data also allowed us to determine if an individual had duplicative coverage, defined as any of the following: purchase of more than one private supplement; purchase of a private supplement in combination with enrollment in a risk-contract HMO; purchase of any additional coverage by Medi-Cal recipients; or purchase of a "dread disease" or hospital indemnity policy in combination with a private supplement or Medicare HMO.

At each survey wave, we asked respondents their current total expenditures on premiums for private supplements or HMOs, or both. Respondents who had coverage that required no premium payment (e. g. a no-premium Medicare HMO) were asked to indicate that their premiums were zero. The analysis reported here focuses on mean premiums for those with coverage requiring a premium.

At waves 1 and 3, we asked participants about their overall satisfaction with their health care coverage and their satisfaction with six different features of their coverage: cost, convenience, continuity, personal treatment, technical quality, and freedom of choice. Each feature, and overall satisfaction, were rated from a low of very dissatisfied (scored as 1) to very satisfied (scored as 5).

SAMPLE CHARACTERISTICS

HIDP participants ranged from 64 to 90 years of age; the mean age was 71.6 years. Reflecting the gender distribution of older persons, 68 percent of workshop participants were female. Nearly half of the participants (46.5 percent) were currently married. Only 12 percent of workshop participants were nonwhite.(8)

The bulk of study participants were retired, although 13 percent reported working part- or full-time for pay. In general, this was a well-educated group, with an average of 13.7 years of education. A rather large group of respondents, 8.4 percent, either did not know their income or did not report it. However, among those who reported their income, the average yearly income was $24,221.(9)

Study participants answered three questions (each scaled from 1 to 4) about their perceived health status (overall rating of health, concern about health, pain level). A response of 1 indicated poor health, a lot of concern about health, a lot of pain; a response of 4 indicated excellent health, no concern about health, no pain. Responses to these questions were summed to form a single health status index, with a higher score indicating better self-reported health status. Overall, participants reported being in reasonably good health, with a mean health status index score of 8.0 (range 3-12).

ANALYSIS

Chi-square tests were used to examine the significance of group differences for categorical variables; t-tests were used to compare group means for continuous variables. Hypothesis testing required that we examine the significance of changes over time for the whole sample, and of differences in changes over time, comparing the intervention and control groups. We used t-test, analysis of variance, and categorical modeling procedures to identify significant time, group, and time-by-group changes.

To determine if randomized assignment had succeeded in achieving group equivalence, we compared members of the intervention and control groups across a wide range of demographic characteristics measured at wave 1. With the exception of age (members of the intervention group were slightly older--72.1 years compared to 71 years, on average), the two groups were statistically similar.

We also compared the baseline demographic characteristics between wave 2 and wave 3 respondents and nonrespondents. These comparisons indicated significant differences in certain key variables such as age, income, ethnicity, health status, and recent hospitalization. Consequently, when examining changes over time in response to the workshop, we limited our analysis to individuals who participated in all surveys relevant to a particular hypothesis.

FINDINGS

CHANGES IN KNOWLEDGE OF MEDICARE

Table 2 presents information on Medicare knowledge before and after the workshop for the whole sample, the intervention group, and the control group. Only participants who answered the questions on both surveys are included in this table (n = 432). For the whole sample, the mean number of correct answers to Medicare questions at wave 1 was 2.78 and at wave 2, 3.59, a mean increase of 0.81 (p < .001). The mean number of "don't know" answers to Medicare questions declined 0.77 from 1.39 at wave 1 to 0.92 at wave 2 (p < .001). These findings confirm our hypothesis that participants in an interactive educational program will demonstrate an increase in their knowledge of Medicare.

[TABULAR DATA 2 OMITTED]

The mean number of correct answers to Medicare questions went from 2.65 at wave 1 to 3.56 at wave 2 in the intervention group; the mean rose from 2.91 to 3.62 in the control group. The slightly greater increase in the mean number of correct answers in the intervention group (0.91 versus 0.71) was not statistically significant. There is a statistically significant difference between the two groups in the decline in "don't know" responses: the intervention group declined more (-0.92) than the control group (-0.63, p < .05). Thus, we have only limited confirmation of our hypothesis that the IEA results in a differentially greater increase in Medicare knowledge.

CHANGES IN KNOWLEDGE OF THEIR OWN INSURANCE COVERAGE

Table 3 presents information on changes in knowledge of their own insurance for the whole sample, including mean correct answers to questions about private supplements and HMOs and standardized scores of own insurance knowledge. The standardized score allows us to assess changes in knowledge regardless of changes in coverage. For the 203 individuals who answered four questions on private supplements at both waves, and about which we were able to determine the correctness of their answers,(10) the mean number of correct answers rose more than half a point, from 1.01 to 1.55 (p < .001). There were 119 individuals who answered questions on HMOs at both waves about which we were able to determine the correctness of their answers.(11) Perhaps because HMO members' knowledge of their coverage was relatively high at baseline, there was only a very slight increase in knowledge of HMOs.

[TABULAR DATA 3 OMITTED]

We had sufficient information about policy characteristics to calculate

a standardized score on own insurance knowledge for 327 individuals who responded to both wave 1 and wave 2.(12) For the sample as a whole, the standardized score rose from 38.16 to nearly 49. This increase of nearly 11 points is statistically significant (p < .001), additional confirmation of our hypothesis that a well- planned educational intervention results in gains in participants' knowledge of their health care coverage.

Table 4 shows that intervention group members demonstrated larger increases in knowledge of private supplements (0.63 points compared to 0.46 points) but smaller increases in knowledge of HMOs (0.02 points compared to 0.11 points). Increases in the standardized score of health insurance knowledge are also greater for the intervention group (11.85 compared to 9.75 percentage points). However, none of these differences are statistically significant at the .05 level, so we find no confirmation of our hypothesis regarding the differential impact of the IEA on gains in own insurance knowledge.

[TABULAR DATA 4 OMITTED]

CHANGES IN HEALTH CARE COVERAGE

Ability to Make Coverage Decisions. Table 5 presents the decision status for the intervention and control groups at wave 3. Note that several respondents indicated both a specific decision (e.g., to stop purchasing a supplement) and that they were still thinking about their coverage. At wave 3, more respondents in the intervention group had purchased new or different supplements or HMOs and had disenrolled from HMOs, but these differences were not statistically significant. Fewer intervention group participants were still thinking about their decision (17.5 percent) than control group members (25 percent; p = .051). This provides only limited support for our hypothesis that IEA workshop participants would be better able to make a decision.

[TABULAR DATA 5 OMITTED]

Kind of Health Care Coverage. Table 6 presents the distribution of coverage for the intervention and control groups at waves 1 and 3, using our four-level categorization (Medicare only; Medicare plus a private supplement; enrollment in a Medicare HMO; and Medicare with a supplement and an HMO). We see the same general pattern of change of coverage in both the intervention and control groups, but in the intervention group more beneficiaries joined a Medicare HMO and fewer have Medicare coverage only.(13) However, these between group differences are not significant at p = < .05.

[TABULAR DATA 6 OMITTED]

Duplicative Coverage. Concerns are often expressed that Medicare beneficiaries purchase duplicative supplemental coverage. Table 7 compares the percent of individuals answering all three surveys, in the intervention and control groups, with duplicative coverage. At wave 1, slight differences between intervention and control groups in the proportion of participants with duplicative coverage were insignificant. By wave 2, the percentage of the intervention group with duplicative coverage had declined to 5.5 percent while the proportion in the control group had increased to 13.4 percent. The difference between the two groups at wave 2 was statistically significant, as was their change from wave 1 to wave 2 (both at p < .01). By wave 3, 9.6 percent of the intervention group had duplicative coverage compared to 17.8 percent of participants in the control group (p < .02). The difference between the two groups in the change from wave 1 to wave 3 is significant (p = .02). These data provide strong confirmation of our hypothesis that the IEA workshops would result in declines in the proportion of participants with duplicative coverage.
Table 7: Intervention and Control Participants
with Duplicative Coverage: Waves 1, 2, and 3
 Significance
 IEA v. Control
Survey Wave Intervention Control P
Wave 1 13.5% 11.9% NS
Wave 2 5.5% 13.4% .007
Wave 3 9.6% 17.8% .016


Premium Expenditures. Participants in the workshops spent less money on premiums after nine months than they did initially, even though we would expect some inflationary rise in premiums over that time period. In the whole sample, premium expenditures declined from $630 at wave 1 to $593 at wave 3 ($37; p < .05). From wave 1 to wave 2, premium expenditures among intervention group participants declined $104, from $679 to $574 per year, while the control participants' mean expenditures rose $17, from $652 to $669. The difference in trend between groups is significant (p = .015). From wave 1 to wave 3, there is a decline in annual premium expenditures in the intervention group of $105, compared to a slight decline of $19 in the control group. Here, the difference in trend approaches, but does not achieve statistical significance at the .05 level (p < .15). These data provide some confirmation of our hypothesis that participants in the IEA workshops would have greater reductions in their premium expenditures.

SATISFACTION WITH HEALTH CARE COVERAGE

Table 8 shows the mean satisfaction scores for the whole sample, the intervention and control groups, and changes in those scores, from wave 1 to wave 3. There were slight but statistically significant increases for the sample as a whole on all features except personal treatment; overall satisfaction stayed steady. The largest increase (+ .36) was found in the cost feature; however, this remained the feature with which respondents were least satisfied. Control group members often showed slightly higher increases in satisfaction than intervention group members, but changes between groups over time were not statistically significant.

[TABULAR DATA 8 OMITTED]

At wave 3, we also asked participants to indicate whether changes they made in their coverage had met their expectations. Table 9 compares the distribution of responses between the intervention and the control groups. A significantly higher proportion of the intervention group reported that the changes they made had met their expectations (15.3 percent versus 5.6 percent; p = .001). This provides limited confirmation of our hypothesis that participation in the Illness Episode workshop would result in an increased satisfaction with health care coverage.

[TABULAR DATA 9 OMITTED]

DISCUSSION AND CONCLUSIONS

Our study represents a rigorous quasi-experimental test of the effects of the IEA and a nonexperimental test of the effects of a carefully planned educational intervention. Our less rigorous test indicates that a three-hour workshop facilitated by trained professionals can significantly increase knowledge of Medicare and their own insurance among beneficiaries who are trying to make a decision and who are sufficiently motivated and functional to participate. Participants in the whole study sample were spending an average of over $50 per year less on premiums, in spite of likely premium increases due to inflation. They indicated significant increases in their satisfaction with almost all features of their health care coverage, especially cost.

When we compare the effects of workshops using the IEA and those using more traditional health insurance information, we see an unexpected pattern of findings. Contrary to our hypotheses, the intervention group did not have significantly greater improvements in insurance knowledge. Nor did they have significantly greater increases in satisfaction with their coverage. Nevertheless, participants in the IEA workshops did show important differences in their purchasing behavior. They were more likely to have made a decision nine months after the workshop, apparently because they were more likely to have changed their coverage.

Perhaps most important, IEA workshop participants were far more likely to have dropped duplicative coverage, while the proportion of the control group with duplicative coverage increased over time. Duplicative coverage was discussed in both workshops, but the IEA workshop, where out-of-pocket costs were a major focus, appears to have emphasized the message that duplicate coverage costs more money in premiums but does nothing to increase reimbursements.(14) Another important finding is that among those who were spending money on premiums, mean expenditures had dropped, after nine months, by over $100 per year. Lower premium expenditures may have been partially related to dropping duplicative coverage. Educational interventions are typically based on the assumption that changes in knowledge will lead to changes in behavior. In the nonexperimental component of this study, we see changes in both knowledge and some behaviors among all workshop participants, regardless of the type of educational intervention. On the other hand, in our quasi-experimental test of the IEA, we found important behavioral changes without differential changes in knowledge. These unexpected findings may be a consequence of insufficiently precise measurement of knowledge: the number of items used was relatively small. It may also be that, given the low baseline knowledge levels in this population, any well-planned educational intervention has such a strong effect that differential effects are swamped.

We are left with the behavioral changes themselves: our findings indicate that participants in the IEA workshops were more likely to make coverage change decisions that indicate greater cost consciousness. Participants in these workshops were not representative of older persons in general: they were better educated and more motivated to seek information relevant to their coverage decisions. Nevertheless, when such a group was exposed to an educational intervention that helped them anticipate the financial consequences of their insurance choices (as called for in social learning theory), and that identified and provided information on key choice criteria (as called for in information processing theory), they not only learned; they used their knowledge to act on their own behalf.

Given this evidence of the positive consequences of insurance education for Medicare beneficiaries, it is encouraging that the Omnibus Budget Reconciliation Act of 1990 provided seed money for state Medigap consumer education programs, and that HCFA and the National Association of Insurance Commissioners are now collaborating regularly on the development and distribution of a Guide to Health Insurance for People with Medicare. OBRA 1990 also legislated useful changes in the marketing of Medigap policies that protect consumers against purchase of duplicate coverage: agents must ask if the consumer has other coverage or is eligible for Medicaid, and must make clear and fair comparisons between an existing policy and one they are offering. The application form must include a statement that the consumer intends to replace an old policy with a new one if one is already held. The legislation also requires an open enrollment period when new Medicare beneficiaries cannot be rejected because of previously existing conditions; limitations for preexisting conditions are also regulated so that they do not apply to those who are changing policies. In addition, OBRA 1990 required that a set of ten standardized Medigap plans be developed. By July 1992, all new Medigap offerings must match one of these ten plans. All insurance companies are required to offer the lowest cost benefit plan; they can also offer additional plans, but except in certain states these offerings cannot vary from those in the standard set. As suggested by consumer information processing theory, this should simplify the selection of Medigap plans. It will be especially valuable when consumers know which benefit plan they want, by facilitating premium comparisons. However, the new legislation still does not ensure that beneficiaries will understand the benefit packages offered and their financial consequences if the beneficiaries become sick. Further, the legislation affect choices among only Medigap plans: it does not affect the more difficult comparisons of Medigap plans with Medicare HMOs or newly emerging Medicare PPOs, as is possible with the IEA. Furthermore, there is still no systematic research being conducted to assess the comparative effectiveness of different approaches to health insurance education. Our findings make clear not only that well-planned education can make a difference, but that different approaches will have different consequences.

NOTES

(1.) Examples are California's Health Insurance Counseling and Advocacy Project (HICAP), Massachusetts' Serving Health Information Needs of Elders (SHINE) program, Washington State's Senior Health Insurance Benefits Advisors (SHIBA) program, and the American Association of Retired Persons' (AARP's) HMO Informed Buyer Project.

(2.) The illnesses selected were hypertension, hearing loss, arthritis, depression, cataract, pneumonia, heart attack, enlarged prostate, diabetes with a complication of gangrene, breast cancer, broken hip with replacement, lung cancer, and stroke.

(3.) A detailed description of the Illness Episode Approach methodology can be found in Sofaer, Davidson, Goodman, et al. 1990. Discussions of the application of the IEA to comparisons of Medicare fee-for-service and Medicare HMOs can be found in Sofaer and Kenney 1989, while its application to Medigap policies is presented in Sofaer and Davidson 1990.

(4.) Copies of the materials used for both intervention and control groups are available on request from the authors. Tables in the articles cited here are similar in content, although not in graphic design, to those used in the workshops.

(5.) Techniques used included brief presentations at senior centers, including various clubs and meals programs; mailings to senior-serving professionals and associations; announcements in senior-oriented newsletters and general press; radio announcements; and posters.

(6.) Spouses, other pairs of relatives, and close friends were assigned to the same workshop to avoid contamination.

(7.) Younger participants included family members and friends of older individuals, as well as people who were interested in getting some advance knowledge of Medicare and other health care options.

(8.) Workshops were offered at senior centers in many but not all parts of the county. A substantially larger number of older persons in Los Angeles County are nonwhite or Latino. Workshops could not be offered in Spanish, so few were held in communities with large proportions of Spanish-speaking residents.

(9.) Years of education and income are calculated based on midpoints of ranges on which data were collected.

(10.) Note that these individuals had supplements of some kind at both waves; this excludes people who switched to or from HMOs or who had no coverage at all.

(11.) Similarly, these individuals had HMOs at both waves, so this excludes people who switched to or from private supplements or who had no coverage at all.

(12.) Individuals with no private coverage at wave 1 are excluded from this analysis, however.

(13.) The nature of the analysis here requires that we use group means rather than comparing individuals who responded to both surveys.

(14.) During the workshops, we suggested that individuals who were enrolling in an HMO for the first time keep their private supplement while they tried out the new option. We gave this advice for two reasons: first, some private supplements have age-related premiums; second, almost all have preexisting-condition limitations. Both of these factors would penalize those who decided to return to fee-for-service Medicare, with a supplement, if dissatisfied with the HMO.

REFERENCES

Acito, F. "Consumer Decision Making and Health Maintenance Organizations: A Review." Medical Care 14, no. 1 (1978): 1-13. Andrews, R. M., B. A. Curbow, E. Owen, and A. Burke. "The Effects of Method of Presenting Health Plan Information on HMO Enrollment by Medicaid Beneficiaries." Health Services Research 24, no. 3 (August 1989): 311-27.

Bandura, A. Social Foundations of Thought and Action. Englewood Cliffs, NJ: Prentice Hall, Inc., 1986. --. Social Learning Theory. 2d ed. Morristown, NJ: General Learning Press, 1977.

Berki, S. E., and M. L. F. Ashcraft. "HMO Enrollment: Who Joins What and Why: A Review of the Literature." Milbank Memorial Fund Quarterly 58, no. 4 (1980): 588-632.

Bettman, J. R. An Information Processing Theory of Consumer Choice. Reading, MA: Addison-Wesley Publishing Company, 1979.

Bettman, J. R., and M. Sujan. "Effects of Framing on Evaluation of Comparable and Noncomparable Alternatives by Expert and Novice Consumers." Journal of Consumer Research 14, no. 4 (1987): 141-54.

Cafferata, G. "Knowledge of Their Health Insurance Coverage by the Elderly." Medical Care 22, no. 9 (1984): 835-47.

Davidson, B. N. "Designing Health Insurance Information for the Medicare Beneficiary: A Policy Synthesis." Health Services Research 23, no. 5 (December 1988): 686-720.

Enthoven, A. "Managed Competition in Health Care and the Unfinished Agenda." Health Care Financing Review Annual Supplement (1986): 105-19.

Farrell, K. "Alternative Health Plans: Informing the Medicare Consumer." Paper presented at the American Public Health Association Annual Meeting, Las Vegas, NV, 1986.

Freidlob, A., and J. P. Hadley. Marketing Medicare in a Competitive Environment. Washington, DC: Department of Health and Human Services, 1985. Garfinkel, S. A., A. J. Bonito, and K. R. McLeroy. "Socioeconomic Factors and Medicare Supplemental Health Insurance." Health Care Financing Review 9, no. 1 (1987): 21-29.

Grazier, K. L., W. C. Richardson, D. P. Martin, and P. Diehr. "Factors Affecting Choice of Health Care Plans." Health Services Research 20, no. 6, part I (1986): 659-82.

Hibbard, J. H., and E. C. Weeks. "Consumerism in Health Care: Prevalence and Predictors." Medical Care 25, no. 11 (1987): 1019-32.

Klowden, J. Ninth Annual Medicare Supplement/HMO Comparison Guide. San Diego, CA: Senior World of California, Spring 1986.

LaTour, S. A., B. Friedman, and E. F. X. Hughes. "Medicare Beneficiary Decision Making about Health Insurance: Implications for a Voucher System." Medical Care 24, no. 7 (1986): 610-14.

Lambert, Z. V. "Elderly Consumers' Knowledge Related to Medigap Protection Needs." Journal of Consumer Affairs 14, no. 2 (1980): 434-51. McCall, N., T. Rice, and J. Sangl. "Consumer Knowledge of Medicare and Supplemental Health Insurance Benefits." Health Services Research 20, no. 6, part 1 (1986): 633-57

Marquis, M. S., D. E. Kanouse, and L. Brodsley. Informing Consumers about Health Care Costs. RAND Publication no. R-3262-HCFA. Santa Monica, CA: The RAND Corporation, 1985.

Mechanic, D. "Consumer Choice among Health Insurance Options., Health Affairs 8, no. 1 (1989): 138-48.

Merrill, J., C. Jackson, and J. Reuter. "Factors That Affect the HMO Enrollment Decision: A Tale of Two Cities." Inquiry 22 (Winter 1985): 388-95.

Shauffler, H. Consumer Education Guidelines for Providers of Health Insurance Information. Contract no. 200-79-0909 by Arthur D. Little, Inc. Atlanta, GA: Bureau of Health Education, Centers for Disease Control, U.S. Department of Health and Human Services, 1980.

Sofaer, S., and B. N. Davidson. "Costs and Benefits of Medigap Insurance: The Illness Episode Approach." Health Care Financing Review 11, no. 4 1990): 121-32.

Sofaer, S., B. N. Davidson, R. D. Goodman, R. Grier, E. Kenney, N. Monk, and A. Ross. "Helping Medicare Beneficiaries Choose Health Insurance: The Illness Episode Approach." The Gerontologist 30, no. 3 (1990): 308-15.

Sofaer, S., and E. Kenney. "Financial Consequences of Joining a Medicare HMO: An Application of the Illness Episode Approach to Estimating Out-of-Pocket Costs." Journal of Health Politics, Policy and Law 14, no. 3 (1989):565-85.

Systemetrics. Disease Staging: A Clinically Based Approach to Measurement of Disease Severity. Vol. 2, Medical Staging Criteria. Washington, DC: U.S. Department of Commerce, National Technical Information Service, 1983a. --. Disease Staging: A Clinically Based Approach to Measurement of Disease Severity. Vol. 3, Appendix C- ICD-9-CM Coded Criteria for Disease Staging. Washington, DC: U.S. Department of Commerce, National Technical Information Service, 1980.

Varner, T., and J. Christy. "Consumer Information Needs in a Competitive Health Care Environment." Health Care Financing Review Annual supplement (1986): 99-104.

Ward, R. A. "HMO Satisfaction and Understanding among Recent Medicare Enrollees." Journal of Health and Social Behavior 28 (December 1987): 401-12.

The research presented in this article was supported by the Health Care Financing Administration (HCFA), under Cooperative Agreement #18-C-98686/0 with the Western Consortium for Public Health. The statements contained herein are solely those of the authors and do not necessarily reflect the views of HCFA.

Address correspondence and requests for reprints to Shoshanna Sofaer, Dr.P.H., Associate Professor, Department of Health Care Sciences, George Washington University Medical Center, 2150 Pennsylvania Avenue, N.W., Room 2B-403, Washington, DC 20037. Erin Kenney, Ph.D. is Project Manager, San Diego State University Foundation, San Diego, CA; and Bruce Davidson, Ph.D. is Program Manager, Health Policy Research Division, SysteMetrics, Santa Barbara, CA. This article, submitted to Health Services Research on September 19, 1990, was revised and accepted for publication on February 7, 1992.
COPYRIGHT 1992 Health Research and Educational Trust
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 1992 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Sofaer, Shoshanna; Kenney, Erin; Davidson, Bruce
Publication:Health Services Research
Date:Dec 1, 1992
Words:6918
Previous Article:How the elderly fare in HMOs: outcomes from the medicare competition demonstrations.
Next Article:Determinants of hospital tax-exempt debt yields: corrections for selection and simultaneous equation bias.
Topics:

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