Alt h ough p olicyma ker s have advocated varying approaches

Transcription

1 Assessing The Impact Of Health Plan Choice Having a choice of health plans is associated with insurance take-up rates, satisfaction with care, and HMO enrollment. by Barbara Steinberg Schone and Philip F. Cooper ABSTRACT: Many health policy researchers have argued that increased insurance plan choice will enhance the efficiency of the health care system. However, relatively little is known about plan choice and its association with insurance coverage and access to and satisfaction with health care. Using data from the 1996 Medical Expenditure Panel Survey, we find that 55 percent of workers had plan choice in that year. Approximately 26 percent of workers with choice obtained it through a family member. Controlling for other factors, plan choice is associated with higher levels of employment-based insurance coverage and a greater likelihood that workers are satisfied that their families health care needs are being met. Alt h ough p olicyma ker s have advocated varying approaches to reforming health care financing, most agree that insurance plan choice is an essential element of a marketbased insurance system. 1 Proponents of managed competition argue that increased consumer choice will promote efficient use of health care resources. It is believed, for example, that increased consumer choice, along with better information and increased priceconsciousness, will reduce health spending and improve quality, as insurers will be forced to compete on the basis of cost and quality rather than on the basis of risk selection. 2 Health plan choice may also allow people to obtain an insurance plan with characteristics they value. For example, people may be able to optimize their coverage for specific types of benefits or obtain access to particular providers or cost-sharing arrangements. If this is so, then we might also expect plan choice to improve satisfaction with and access to health care. Moreover, greater plan choice may result in higher insurance take-up rates. Because it appears likely that our nation will retain its current employer-based coverage system, we need to improve our understanding of the availability of plan choice. The purpose of this paper is to investigate the prevalence of plan choice and the relationship DATAWATCH 267 Barbara Schone and Philip Cooper are senior economists in the Division of Social and Economic Research, Center for Cost and Financing Studies, at the Agency for Healthcare Research and Quality (AHRQ) ProjectHOPE ThePeople-to-People Health Foundation, Inc.

2 D a t a W a t c h 268 HEALTH PLAN CHOICE between plan choice and several outcomes (coverage, access, satisfaction, and managed care enrollment). We take a broad view of plan choice and specifically account for the availability of employment-based insurance from other family members. 3 Previous studies indicate that a large proportion of workers are not offered a choice of health plans. 4 With the exception of research using Community Tracking Study data, most studies have focused solely on health insurance choices available in the workplace. 5 Because the number of dual-earner couples is rising, some workers who are not offered a choice of plans from their own jobs may get choice through a spouse. 6 Our study relies on data from the household component of the Medical Expenditure Panel Survey (MEPS-HC), which represents an alternative source of data on plan choice (and its sources) that has not yet been analyzed. We analyze the association between plan choice and other health insurance related outcomes, including insurance coverage and satisfaction. We also include uninsured persons in the analysis, which enables us to assess whether there is a relationship between choice of plans and insurance coverage. 7 If decisions to take up employers offers of insurance seem to be related to the availability of plan choice, then choice may both encourage enrollment and improve satisfaction with health care. Data And Methods We use a nationally representative sample of wage earners (those who are not self-employed) ages from Round 1 of the 1996 MEPS to conduct our analysis. 8 Our sample totals 8,594 persons, representing roughly 100 million workers in The main variable of interest for our analysis is whether a worker had access to more than one health insurance plan. Although the tenets of managed competition highlight the importance of offering workers a choice of health plans, many workers indirectly have a choice of plans through a family member. To account for these alternative ways of obtaining choice, we determined whether workers were offered a choice of plans through their own jobs or indirectly. 9 We classified workers along three avenues: (1) they were offered a choice of plans through their own jobs; (2) they were offered a single plan through their own jobs and had a family member (usually a spouse) who was offered one or more plans; or (3) they were not offered insurance but had a family member who was offered a choice of plans. 10 Although the data are now several years old, they currently represent the most comprehensive source of household-based data on health insurance and health-related information. In addition to basic demographic, employment, health insurance H E A L T H A F F A I R S ~ V o l u m e 2 0, N u m b e r 1

3 status, and geographic information available in the Round 1 data, we include information related to health care access and satisfaction from Round 2 of the survey. Access to care is represented by having a usual source of care. Satisfaction is measured by an indicator that identifies workers who are very satisfied that their families health care needs are being met. 11 We also used multivariate regression analyses to assess the relationship between plan choice and health insurance coverage, access to care, and satisfaction. Our regression models include controls for basic demographic and employment characteristics, and we report the marginal effects of having plan choice on the outcomes of interest, holding these other factors constant. All results discussed below are statistically significant at p <.05 unless otherwise noted, and standard errors are adjusted for the complex survey design. Insurance And Plan Choice Availability Approximately 41 percent of workers were offered a choice of health plans from their own jobs in 1996, with the remainder either not being offered insurance or being offered only one plan (Exhibit 1). Among workers who were offered health insurance, almost 55 percent had a choice of plans. These findings are roughly similar to choice rates found in establishment-based studies and are remarkably similar to choice estimates from the Community Tracking Study. 12 Thus, our findings, like those of others, indicate that a significant fraction of workers lack access to a choice of plans from their own jobs. Taking into account insurance availability through family members increases the prevalence of plan choice among workers by DATAWATCH 269 EXHIBIT 1 Distribution Of Workers By Health Insurance And Plan Choice Availability, 1996 Workers access to choice (from own job) Not offered insurance Offered 1 plan Offered a choice of plans Workers access to choice (from any source) No access to employment-based insurance Access to 1 employment-based plan Access to at least 2 employment-based plans Source of choice Own job Spouse offered at least 2 employment-based plans Workers and spouse each offered 1 plan % % SOURCE: Round 1, 1996 Medical Expenditure Panel Survey (MEPS), Agency for Healthcare Research and Quality. a Not applicable. a a a

4 D a t a W a t c h 270 HEALTH PLAN CHOICE roughly fourteen percentage points. 13 On net, one-fourth of workers who have a choice of plans obtain it through a family member (Exhibit 1). Thus, insurance options available through family members have an important impact on the choices available to workers, although a sizable fraction of workers still lack choice. n Insurance-related outcomes. The data also reveal a relationship between access to insurance, plan choice, and all of the outcomes we consider. For example, we observe that workers without access to employment-based coverage had the highest rates of being uninsured, while workers with choice had the lowest rates (Exhibit 2). Similarly, the likelihood of taking up coverage from a worker s own job was highest for workers with choice, as was the rate of job-based coverage, as reflected in the family take-up rate, for workers with access to at least two insurance plans. Usual source of care and satisfaction. The bivariate findings also indicate that regardless of our measure of plan choice, the likelihood of having a usual source of care was highest among those workers with a choice of plans and lowest for workers without access to employment-based coverage. Similar differences in our satisfaction measure are also observed, although it is interesting to note that the proportion of workers who were very satisfied with health care was quite high for all workers. 14 HMO enrollment. We also found that workers with private insurance coverage and a choice of plans were significantly more likely than workers with access to only one plan (or not offered coverage) were to be enrolled in a health maintenance organization (HMO). EXHIBIT 2 Insurance Status And Take-Up Rates, Access To And Satisfaction With Care, And HMO Enrollment, By Health Insurance And Plan Choice Availability, 1996 Workers access to choice (from own job) Not offered insurance Offered 1 plan Offered a choice of plans Workers access to choice (from any source) No access to employment-based insurance Access to 1 employment-based plan Access to at least 2 employment-based plans 48.6% d 76.9% 82.8 d % d % % SOURCE: Round 1, 1996 Medical Expenditure Panel Survey (MEPS), Agency for Healthcare Research and Quality. NOTES: The data are based on full-year respondents only and are reweighted accordingly to account for nonresponse. HMO is health maintenance organization. a Among those offered insurance from their own job. b Proportion of workers with employment-based coverage among those with access from any family source. c Among those workers with private health insurance coverage. d Not applicable % d H E A L T H A F F A I R S ~ V o l u m e 2 0, N u m b e r 1

5 This may reflect the fact that larger firms that offer insurance must offer an HMO option if there is one available. 15 Other characteristics. These findings may reflect other underlying differences in the characteristics of workers, as well. Overall, we find significant differences in the characteristics of workers across the health insurance and choice categories (Exhibit 3). Workers who had a choice of insurance plans from their own jobs were least likely to be employed in the private sector, to be employed in small establishments, and to be low-wage workers (less than $7 per hour in 1996). Similarly, these workers were most likely to work full time, be a union member, and have other fringe benefits. As expected, the employment characteristics of workers who obtained a choice of insurance plans only through a family member appeared to be most similar to workers who were offered only a single plan. We also found important differences in sociodemographic characteristics. For example, Hispanic workers were disproportionately overrepresented among workers without access to job-based insurance and were underrepresented among workers with a choice of plans (overall, 9.6 percent of workers in 1996 were Hispanic, and only 6.7 percent of these workers had a choice of insurance plans from their own jobs). Also, workers with access to a choice of plans had the highest socioeconomic status (Exhibit 3). Interestingly, the likelihood of living in an urban area was similar for all groups of workers, except those who had plan choice from their own jobs. This may indicate that there are supply-side impediments to providing choice in rural areas since it may be difficult to offer two plans DATAWATCH 271 EXHIBIT 3 Characteristics Of Workers, By Health Insurance And Plan Choice Availability, 1996 Related to employment Works in small establishment Works in private sector Has retirement benefits Has paid vacation Has sick pay Is a union member Is employed full time Is a low-wage worker Sociodemographic Years of education Hispanic Black Poor or near-poor High income Young Lives in urban area 55.3% % % % % % % % SOURCE: 1996 Medical Expenditure Panel Survey (MEPS), Agency for Healthcare Research and Quality. NOTES: Poor or near-poor indicates income less than 125 percent of poverty. High income refers to income more than 400 percent of poverty. Young workers are under age twenty-five. Low-wage workers are those earning less than $7 per hour.

6 D a t a W a t c h 272 HEALTH PLAN CHOICE that are able to compete on the basis of cost and quality. n Predicted probabilities for each outcome. Given the underlying differences in workers characteristics, we examine plan choice more carefully by focusing on workers with access to health insurance and using multivariate analysis. Specifically, we estimate probit regressions that include controls for urbanicity, region of residence, employment characteristics (wage rate, union membership, industry, firm size, working in the private sector, and whether the worker has retirement, sick pay, and paid vacation benefits), and sociodemographic characteristics (age, sex, marital status, education, race/ethnicity, health status, and poverty status) to determine whether the association between choice and outcomes remains once other differences are taken into account. 16 The results indicate that once other factors are taken into account, plan choice is less strongly related to the outcomes we consider. For example, we no longer find that choice is associated with being uninsured or taking up coverage from one s own job (Exhibit 4). Plan choice is associated, however, with an increase in the likelihood of having employment-based coverage from any source, of workers being very satisfied that their families health care needs are being met, and of having a usual source of care (although the underlying coefficient here is not statistically significant at conventional levels). Plan choice also is strongly related to HMO enrollment, which is not surprising since firms that offer choices are more likely to offer a managed care plan. 17 Corresponding differences from Exhibit 1 are much larger, which suggests the importance of other factors in explaining differences in the outcomes we consider. EXHIBIT 4 Predicted Probabilities Of Health Insurance Related Outcomes, By Plan Choice Predicted probability Choice available No choice available Net marginal (percentage-point) effect of having plan choice ** ** 16.6*** SOURCE: 1996 Medical Expenditure Panel Survey (MEPS), Agency for Healthcare Research and Quality. NOTES: The results are based on probit regressions on workers that have access to employment-based coverage. They control for worker characteristics (age, sex, race/ethnicity, poverty status, health status, region of residence, education, and marital status) and employment characteristics (wage, establishment size, union membership, industry, a private-sector indicator, and the availability of a retirement plan, sick pay, and paid vacation). HMO is health maintenance organization. a Regression based on workers offered insurance from their own jobs. For the take-up regressions, choice of plans refers to choice from worker s own job; take-up equals 1 if the worker holds coverage from his own job (0 otherwise). b Family take-up equals 1 if worker is covered by job-based insurance; 0 otherwise. c Based only on workers who are married. d For workers who have private health insurance coverage. ** p <.05 *** p < H E A L T H A F F A I R S ~ V o l u m e 2 0, N u m b e r 1

7 Discussion And Policy Implications We find that 59 percent of workers (representing approximately fifty-nine million) did not have access to a choice of health insurance plans from their own jobs in 1996; these findings are consistent with the results of other studies. While accounting for insurance options available through other family members gave an additional fourteen million workers access to multiple plans, 45 percent of all workers still lacked choice. To the extent that choice is a fundamental component of a market-based health insurance system, identifying ways to improve choices to workers requires further attention. Our results provide mixed evidence on the relationship between having a choice of plans and insurance coverage, access, and satisfaction. The finding of a positive, albeit modest, relationship here suggests that the advantages of plan choice may extend beyond the efficiency advantages ascribed to it by the proponents of managed competition. However, the relatively high rates of health care satisfaction and access among all workers to whom insurance was available suggest that firms that offer only a single plan may be doing an adequate job of satisfying their employees preferences. Our results also indicate that workers in small establishments and in rural areas were more likely to be offered only a single plan. These geographic variations suggest that there may be some areas in which it will be difficult to enhance plan choice. 18 Targeting efforts at increasing insurance coverage for rural residents and finding effective ways of sustaining multiple plans clearly need policy attention if managed competition is to be viable throughout the entire United States. Insurance-purchasing pools may be one mechanism for improving plan choice, although the effects appear to be modest. 19 Another alternative, proposed by the American Association of Health Plans, is to offer tax incentives to firms that offer multiple plans. 20 Additional research on the likely success of alternative mechanisms to enhance plan choice is clearly needed. 21 DATAWATCH 273 The views in this paper are our own. No official endorsement by the Agency for Healthcare Research and Quality or the Department of Health and Human Services is intended or should be inferred. An earlier version of this paper was presented at the 1999 Association for Health Services Research (AHSR) meetings in Chicago. We thank Steve Cohen, Don Metz, Alan Monheit, Eric Schone, and two anonymous reviewers for helpful comments. NOTES 1. See, for example, H.J. Aaron, Serious and Unstable Condition: Financing America s Health Care (Washington: Brookings Institution, 1991); and A.C. Enthoven and S.J. Singer, Market-Based Reform: What to Regulate and by Whom, Health Affairs (Spring 1995):

8 D a t a W a t c h 274 HEALTH PLAN CHOICE 2. See P. Ellwood, A. Enthoven, and L. Etheredge, The Jackson Hole Initiatives for a Twenty-first Century American Health Care System, Health Economics (October 1992): One concern, however, is that plan choice may result in market segmentation, exacerbating adverse selection. See D. Cutler and R. Zeckhauser, The Anatomy of Health Insurance, NBER Working Paper no (Cambridge, Mass.: National Bureau of Economic Research, 1999). 3. In this analysis we focus on employment-based plan choice and the choices that were available to workers in Most workers, however, implicitly have plan choice. Workers without employment-based plan choice through their current jobs could obtain access to additional insurance plans by purchasing insurance in the individual market, obtaining eligibility for public coverage, or changing jobs. Such options, however, are generally costly. 4. See, for example, Center for Studying Health System Change, How Widespread Is Managed Competition? Data Bulletin no. 12 (Washington: HSC, Summer 1998); K. Davis et al., Choice Matters: Enrollees Views of Their Health Plans, Health Affairs (Summer 1995): ; R. Ullman et al., Satisfaction and Choice: A View from the Plans, Health Affairs (May/June 1997): ; and A.A. Gawande et al., Does Dissatisfaction with Health Plans Stem from Having No Choices? Health Affairs (Sep/Oct 1998): As in our research, S. Trude, Center for Studying Health System Change, Who Has a Choice of Health Plans? Issue Brief no. 27 (Washington: HSC, February 2000), distinguishes between choice from a worker s own job and through a family member. Gawande and colleagues definition of plan choice refers to any plan choice but does not provide information on the source of choice. 6. There has been speculation that some firms make coverage available only to employees. However, data from the 1997 Robert Wood Johnson (RWJ) Employer Health Insurance Survey indicate that approximately 98 percent of employees that are offered insurance have access to dependent coverage as well. See S. Long and M.S. Marquis, Trends in Offering Employer-Sponsored Coverage, Center for Studying Health System Change Data Bulletin no. 15 (Washington: HSC, Fall 1998). Although less common, it is possible that a working child without a choice of plans could obtain such a choice from a parent. This situation could occur if the child could still be covered as a dependent on his parents insurance policy. 7. The uninsured population is of interest since an increasing number of uninsured persons have access to employment-sponsored health insurance. See P.F. Cooper and B.S. Schone, More Offers, Fewer Takers for Employment- Based Health Insurance: 1987 and 1996, Health Affairs (Nov/Dec 1997): MEPS provides detailed information on health insurance and health care use and expenditures, along with basic socioeconomic information for a nationally representative sample of the civilian, noninstitutionalized U.S. population. See J. Cohen et al., The Medical Expenditure Panel Survey: A National Health Information Resource, Inquiry (Winter 1996/1997): Information about the types of choices available is unavailable in the household component of MEPS. All workers were asked if they were offered insurance from their jobs and, if so, whether a choice of plans was available. Moreover, this information was ascertained regardless of insurance status. 10. We assumed that choice through a spouse was not possible if the worker and spouse had the same employer. 11. Relative to workers who are somewhat satisfied, or somewhat or very dissatisfied, that their families health care needs are being met. 12. For example, the 1997 RWJ employer survey indicates that 41 percent of workers offered insurance had a choice of plans. The Community Tracking Study, a household-based survey, found that 53 percent of workers offered H E A L T H A F F A I R S ~ V o l u m e 2 0, N u m b e r 1

9 insurance had a choice of plans. 13. Our estimate of the number of persons with insurance through a family member may be slightly underestimated because we consider only the insurance that is available through others in the same household. Some persons (for example, divorced persons) may also obtain insurance access from someone outside the household, which the MEPS data do not capture. 14. Since information on access and satisfaction comes from later rounds of the survey, there is some potential for attrition bias in the results. To address this concern, the results in Exhibit 2 are based on full-year respondents and are reweighted accordingly to account for nonresponse. 15. See 42 U.S. Code Section 300e We believe that our multivariate results are best interpreted as correlations that are corrected for observable variables rather than a causal relationship. First, plan choice is likely to be correlated with a number of unobservable variables including plan attributes (copayments, premiums, and overall plan generosity). Since data are unavailable on such additional measures, it is unclear whether choice per se is related to the outcome of interest (or whether it simply is reflecting these other underlying differences in plan generosity). Moreover, to establish a causal relationship, plan choice should be modeled as an endogenous variable, since workers may sort across firms according to insurance preferences (see A. Monheit and J. Vistnes, Health Insurance Availability at the Workplace How Important Are Worker Preferences? Journal of Human Resources 34, no. 4 (1999): for evidence of this type of behavior), and employers may formulate their insurance plans (including the provision of choice) strategically. See D. Dranove, K. Spier, and L. Baker, Competition among Employers Offering Health Insurance, Journal of Health Economics ( January 2000): ; and J. Moran, M. Chernew, and R. Hirth, Preference Diversity and the Breadth of Employee Health Insurance Options, Health Services Research (forthcoming). 17. Firms that offer a single plan may be less inclined to offer an HMO because it may be more difficult to satisfy diverse worker preferences with such a plan. Data reported by J.C. Cantor, S.H. Long, and M.S. Marquis, Private Employment-Based Health Insurance in Ten States, Health Affairs (Summer 1995): , from the 1993 RWJ Employer Health Insurance Survey, indicate a correlation between having a choice of plans and having an HMO option. 18. See R. Kronick et al., The Demographic Limitations of Managed Competition, New England Journal of Medicine (14 January 1993): See S.H. Long and M.S. Marquis, Pooled Purchasing: Who Are the Players? Health Affairs (July/Aug 1999): ; and S.H. Long and M.S. Marquis, Have Small-Group Health Insurance Purchasing Alliances Increased Coverage? Health Affairs (Jan/Feb 2001): See Improving Access to Health Insurance: Now and in the Future (Washington: American Association of Health Plans, February 2000). 21. A thorough understanding of plan choice requires data from multiple years to understand how the dynamic nature of the health insurance market affects workers perceptions of plan choice. According to J. Gabel et al., Job-Based Health Insurance in 2000: Premiums Rise Sharply while Coverage Grows, Health Affairs (Sep/Oct 2000): , there has been a large increase in pointof-service (POS) and preferred provider organization (PPO) coverage relative to conventional coverage between 1996 and 2000 (HMO coverage has remained stable). Such changes may have changed access to providers and, as a result, affected workers demands for plan choice. DATAWATCH 275