Do HMOs Make a Difference?

Originally published by the Center for Studying Health System Change (HSC). HSC was a nonpartisan policy research organization funded principally by the Robert Wood Johnson Foundation.

Comparing Access, Service Use and Satisfaction Between Consumers in HMOs and Non-HMOs

James D. Reschovsky, Peter Kemper, Ha T. Tu, Timothy K. Lake, Holly L. Wong

As HMO enrollment surged through the 1990s -- eventually covering more than half of all privately insured individuals -- questions about the impact on consumers multiplied. This Issue Brief presents findings from a large national study of the privately insured population under age 65. The research found no detectable difference between HMOs and other insurance types in the use of three costly services: inpatient hospital care, emergency room visits, and surgeries. Reports of unmet need or delayed care were also negligible between the two groups. Where differences did appear, they presented a trade-off: HMOs delivered more primary and preventive services and reduced financial barriers to care, but they provided less specialist care and created more administrative barriers. Patients in HMOs also reported lower satisfaction, less trust in physicians, and lower ratings of doctor visits.

The Policy Debate over HMOs

HMOs had been credited with bringing health care cost growth under control during the 1990s, but both consumers and providers were raising concerns about whether cost savings came at the expense of care quality. Public anxiety about restricted provider choice, limited services, and the quality of care fueled a managed care backlash and generated support for government regulation of HMOs.

Industry defenders countered that negative press coverage was based on unrepresentative cases while the typical HMO experience went unreported. They argued that HMOs' emphasis on prevention and care management produced better, more cost-effective care than less managed forms of insurance.

Much of the debate, however, was hampered by a shortage of current, broad-based empirical evidence. This study aimed to narrow that gap by systematically comparing how HMOs affected access, service use, and patient satisfaction across a nationally representative sample.

Analyzing HMO Effects

Previous research comparing HMO and non-HMO care had yielded inconsistent results, often because of methodological limitations: studies relied on outdated data, examined narrow populations such as enrollees of a few plans or residents of a single city, and used varying approaches to control for differences in who enrolled in HMOs versus other coverage. This study addressed those shortcomings by drawing on a large, nationally representative sample of privately insured individuals under 65, using data collected in 1996 and 1997.

The analysis covered multiple dimensions. Access to care was measured through reports of unmet need and the presence of financial or administrative barriers to obtaining care. Service use encompassed ambulatory visits of various types, emergency room use, hospital stays, surgeries, and preventive services such as mammograms. Consumer assessments included overall satisfaction with health care, satisfaction with physician choice, ratings of doctor visits, and measures of patient trust in physicians.

Crucially, the researchers controlled for a wide range of enrollee characteristics -- health status, income, demographics, and other factors -- to avoid the bias that would result from simply comparing raw outcomes between the two insurance types.

Little Difference on Key Measures

The study found no evidence of meaningful differences between HMO enrollees and those in other insurance types in hospital use, emergency room visits, or rates of surgery. Reports of unmet care needs or delayed care differed only slightly.

The difference in hospital days between HMO and non-HMO enrollees was small and not statistically significant. Overall surgery rates did not differ significantly either, and there was no evidence that HMOs shifted surgery from inpatient to outpatient settings. HMO enrollees used emergency rooms at comparable rates to those in other plans. Among the costly services examined, only specialty care showed a significant difference, with HMO enrollees using specialist services less frequently.

These findings ran counter to conventional wisdom about where HMOs had cut costs. Two explanations were offered. First, if rising HMO market share had changed how all providers treated patients -- a spillover effect -- then the comparison between HMO and non-HMO care at the individual level would naturally narrow. Second, HMOs and other products had been converging: traditional fee-for-service and PPO plans had adopted managed care tools like preadmission authorization, while HMOs had responded to competitive pressure by broadening their networks and offering out-of-network coverage.

While these results offered some reassurance that HMOs were not stinting on costly but necessary care, they were less encouraging for those who saw HMOs as the vehicle for controlling spending by reducing inappropriate care. HMOs had achieved at least some of their cost containment through negotiated provider discounts. Whether they could sustain cost control through more sophisticated care management -- rather than simply negotiated lower prices -- was an open question, particularly given the slow implementation of advanced care management tools and the information systems needed to support them.

For unmet need and delayed care, HMO enrollees were only marginally more likely (4.4 percent vs. 3.7 percent) to report that they did not receive care they felt they needed, a small but statistically significant difference. Reports of delayed care showed no difference at all (12.8 percent vs. 12.7 percent). There was also no evidence of differences in appointment wait times, travel time to care, or office waiting times.

Trade-Offs on Other Dimensions

Where HMO effects did show up, they presented a mixed picture. On the positive side, HMO enrollees had more primary care visits, received more preventive services such as mammograms, and faced lower financial barriers to care because of reduced cost-sharing. On the negative side, they had fewer specialist visits, encountered more administrative barriers when seeking care, and reported lower satisfaction with their health care, less trust in their physicians, and lower ratings of their most recent doctor visit.

This trade-off was central to the ongoing policy debate. HMO proponents pointed to the gains in primary care and prevention as evidence that managed care was redirecting the system toward more appropriate care patterns. Critics focused on the reductions in specialty access and the consistently lower satisfaction scores as signs that HMOs were compromising the patient experience.

Implications of the Findings

The results suggested that the most alarming claims about HMOs -- that they were systematically denying needed hospital care and emergency services -- were not supported by the data for the typical enrollee. At the same time, the evidence did not support the most optimistic claims either: HMOs were not demonstrating clearly superior efficiency in the use of costly services at the individual level.

The lower satisfaction and trust scores in HMOs were notable. Research had shown that consumers who had a choice of plans were more satisfied regardless of plan type, suggesting that the managed care backlash may have stemmed partly from consumers feeling locked into plans they did not choose. The findings reinforced the case for policies that expanded consumer choice and ensured that plan restrictions were transparent.

The study could not assess the clinical appropriateness or quality of care within the services measured. The number of hospital days, surgeries, or specialist visits might be the same, but the content and quality of those encounters could differ. Additional research using medical records and clinical quality measures would be needed to answer those questions.

Data Sources and Methodology

The analysis used data from HSC's Community Tracking Study (CTS) Household Survey, conducted in 1996-1997, which included a large, nationally representative sample of the civilian, noninstitutionalized population. The study focused on privately insured individuals under age 65. Multivariate statistical methods controlled for health status, demographics, income, and other factors to isolate the effects of HMO enrollment from the characteristics of the people who selected HMOs.

Sources and Further Reading

HSC Community Tracking Study Household Survey, 1996-1997.

Related HSC publications: "Managed Care Woes," Issue Brief No. 13, May 1998; "Health Care Perceptions and Experiences," Issue Brief No. 30, September 2000; "Who Is Likely to Switch Health Plans?" Data Bulletin No. 18, July 2000.