This information is for reference purposes only. It was current when produced and may now be outdated. Archive material is no longer maintained, and some links may not work. Persons with disabilities having difficulty accessing this information should contact us at: https://info.ahrq.gov. Let us know the nature of the problem, the Web address of what you want, and your contact information.
Please go to www.ahrq.gov for current information.
Hospitalization within 10 days of outpatient treatment for pneumonia may
indicate less than optimal care
Due to the growing influence of managed care organizations on health care delivery, it is likely
that an increasing number of patients with pneumonia will be treated as outpatients. To assure
quality of care for these patients, regular monitoring of patient outcomes will be critical. A key
indicator of appropriateness of care is the hospitalization rate of patients treated in the outpatient
setting, particularly those hospitalized within 10 days of initial treatment. While a small percentage
of patients at low risk for pneumonia complications can be expected to end up in the hospital, a
higher than usual proportion of subsequent hospitalizations for pneumonia complications may be a
signal that quality of care could be compromised.
Led by Wishwa N. Kapoor, M.D., M.P.H., of the University of Pittsburgh, and supported by the
Agency for Health Care Policy and Research (HS06468), the Pneumonia Patient Outcomes
Research Team (PORT) prospectively studied 944 adults treated initially at outpatient sites
participating in a multicenter study of community-acquired pneumonia (CAP). They found that 90
percent of all CAP-related hospitalizations occurred within 10 days of the initial outpatient visit.
On the other hand, 81 percent of hospitalizations that occurred more than 10 days later were
related to coexisting illnesses, such as cardiac disease or chronic obstructive pulmonary disease.
Outpatients who were subsequently hospitalized took longer to return to usual activities (a
median of 14 vs. 6 days) and had a higher 30-day mortality rate than those who were not
hospitalized (4.3 percent vs. .3 percent). This difference in medical outcomes may be explained in
part by increased age, more comorbid illness, and higher levels of severity of illness among
outpatients who are subsequently hospitalized.
Details are in "Patients hospitalized after initial outpatient treatment for community-acquired
pneumonia," by Michael F. Minogue, M.D., M.S., Christopher M. Coley, M.D., Michael J. Fine,
M.D., M.Sc., and others, in the March 1998 Annals of Emergency Medicine 31(3), pp.
Clinical guidelines must be implemented with other approaches to
substantially improve quality of care
Pediatric clinical practice guidelines originate largely from professional societies and Federal or
State agencies, as well as managed care organizations. About 28 States require the adoption of
guidelines in their contracts with health plans serving Medicaid patients. Strong evidence exists
that well-implemented guidelines can be used to achieve improvements in both the processes and
outcomes of care. However, the magnitude of improvement varies widely. Also, mailing
guidelines alone, like any other single intervention, is unlikely to improve the quality of care
Guidelines must be implemented in combination with other approaches, such as local clinical
leadership, education, reminders, and incentives, assert Lisa Simpson, M.B., B.C.H., M.P.H.,
Douglas Kamerow, M.D., M.P.H., and Irene Fraser, Ph.D., of the Agency for Health Care Policy
and Research. Their review of current literature showed that other factors affecting the
implementation of guidelines include the consistency of guideline information with current
opinions and practices, the skills and resources necessary to implement knowledge, financial
incentives, reimbursement policies, patient preferences, perceived implications for litigation,
extent of physician participation in the development or adaptation of the guideline, and physician
perception of the philosophical underpinnings of the guideline (for instance, whether the incentive
was to restrain costs or improve quality of care).
There is evidence that guidelines may not always be having their desired impact on pediatric care.
For instance, one report showed that immunization rates for children enrolled in managed care
plans varied from 59 to 81 percent. A number of studies directly evaluating the cost impact of
guideline implementation have found varying degrees of savings. Despite this variation in
guideline impact, guidelines will continue to grow as part of the pediatric landscape, note the
authors. They conclude that practice guidelines can improve the quality of care when implemented
appropriately, but physicians must be involved in their development and implementation.
See "Pediatric guidelines and managed care: Who is using what and what difference does it
make?" by Drs. Simpson, Kamerow, and Fraser, in the April 1998 Pediatric Annals 27(4),
Reprints (AHCPR Publication No. 98-R081) are available from the AHCPR Publications Clearinghouse.
Return to Contents
Today's primary care physicians are being asked to approach their
practices with a community perspective
Primary care physicians (PCPs) are increasingly being asked to step out of their offices and into
the neighborhoods where their patients live and work to help solve some of the health problems
facing their communities. Apparently, PCPs are involved in community-oriented primary care
(COPC) in some ways but not others. And surprisingly, physicians who work with poor and
minority groups are less involved in their communities than other physicians, according to a study
supported in part by the Agency for Health Care Policy and Research (National Research Service
Award Training grant T32 HS00032).
University of North Carolina researchers mailed a questionnaire to a random sample of 500 young
PCPs (medical school graduation from 1985 to 1990) in the United States (247 responded),
asking them how confident they were in performing each of 15 specific community-relevant
activities that comprised four areas of COPC. Nearly two-thirds of PCPs had spoken to a
community group about a health issue within the previous 2 years. However, only one-third had
worked with a community group to address a local health problem, a key component of COPC.
Physicians' attention to sociocultural aspects of patient care centered mostly around inquiries
about the impact of illness on patients' lives and about their social supports. They inquired least
often about patients' transportation problems and use of unconventional medical treatments,
which is common in some communities.
Physicians were most familiar with community health resources such as physical therapists, social
workers, and nutritionists. They were least familiar with the less traditional or mainstream
resources such as local chiropractors, women's shelters, and clergy. Physicians generally felt
assimilated into their communities in terms of acceptance, feeling appreciated and having valued
friends, but indicated they were not active in community organizations.
The researchers call for future studies to examine what insurers and practice organizations that
provide care for needy populations can do to encourage and promote community involvement
among their participating physicians.
For more details, see "The four community dimensions of primary care practice," by Donald E.
Pathman, M.D., M.P.H., Beat D. Steiner, M.D., M.P.H., Eric Williams, Ph.D., and Toija Riggins,
Ph.D., in the April 1998 Journal of Family Practice 46(4), pp. 293-303.
Probable practice location and patient mix should inform medical
students' primary care residency decisions
Medical students considering primary care practices can choose from family medicine, internal
medicine, or general medicine. Their choice should take into consideration the probable location
of their practice and the scope of patient problems they are likely to see, note Carolyn M. Clancy,
M.D., and James K. Cooper, M.D., of the Agency for Health Care Policy and Research.
In a recent commentary, Drs. Clancy and Cooper highlight the history and current practice
patterns of these three specialties. They point out that family physicians clearly have the potential
for the broadest spectrum of practice, since they are trained to provide care for individuals from
conception to death, including prenatal and obstetric care. In contrast, internists and general
medicine-pediatricians spend more time training in the hospital, and their practices include patients
with a greater burden of illness. Overall, internists have older patients with a higher number of
chronic conditions than family physicians. Also, medicine-pediatric physicians tend to see more
children under 2 years of age and also patients with more complex illnesses than their family
Several trends will affect all physicians' practices in the near future: an aging population, a
growing number of managed care patients, and more large group practices with a corresponding
need for clinical leadership. Also in the future, primary care physicians (PCPs) will coordinate and
possibly deliver care during and after patient transitions among varying levels of care, such as
home care and long-term care settings. Some PCPs (mostly internists) may become hospitalists,
who devote their time to hospitalized patients only, as has long been done in many Western
European countries. Drs. Clancy and Cooper predict that increased attention to community health
will be a defining characteristic of primary care practice in the 21st century.
See "Approaches to primary care: Current realities and future visions," by Drs. Clancy and
Cooper, in the March 1998 American Journal of Medicine 104, pp. 215-218.
Reprints (AHCPR Publication No. 98-R051) are available from the AHCPR Publications Clearinghouse.
Return to Contents
Workshop highlights pros and cons of carve-out contracts for specialty
So-called "carve outs" are one strategy used by health plans, employers, and public payers to
contract with specialty providers to care for particular groups of patients. For instance, they may
carve out or contract with an exclusive group of providers to manage clients with mental health
problems, AIDS, or cancer. In addition, contracts may be used for a set of surgical procedures,
such as those dealing with cardiovascular disease.
Theoretically, carve-out organizations can select and assemble the best specialists; treat large
numbers of patients with the same condition(s); adopt and evaluate the newest technologies,
drugs, and clinical guidelines; and conduct research. On the positive side, carve outs within some
employee benefit programs reduce the opportunities for risk selection—that is, attempts by
health plans to avoid enrolling individuals with conditions that are costly to treat.
On the other hand, problems with administrative functions (payment and oversight) can occur
because carve-out services are provided outside the usual network of providers. Also, problems in
coordinating care received by patients in a carve-out arrangement with the care provided by their
primary care physicians may reduce the overall quality of care.
In January 1998, the Agency for Health Care Policy and Research sponsored a workshop to
define what is and is not known about carve outs and to develop a research agenda on this topic.
According to evidence presented at the workshop, carve outs can lower health care costs,
particularly in the area of mental health. However, conference participants and attendees noted
that we need more research about the effects of carve outs on quality and access, according to
AHCPR researchers Bernard Friedman, Ph.D., Kelly Devers, Ph.D., Fred Hellinger, Ph.D., and
Irene Fraser, Ph.D. They are the authors of an overview article on carve outs and similar
arrangements which appears in a special issue of the American Journal of Managed Care. The
authors concluded that there are serious concerns about the impact of carve outs and that
additional rigorous evidence is needed about how carve outs affect quality and access. They also
noted, however, that this type of research is difficult to conduct because of the proprietary nature
of the data collected and maintained by carve-out organizations.
The special issue also contains five commissioned papers originally presented at the workshop.
Taken together, these five papers provide an excellent synthesis of what is known about the array
of carve-out and specialty contracting models that exist today. They also provide a firm
foundation for building a research agenda. The papers are:
- "The economic functions of carve outs in managed care," (pp. SP31-39) by Richard G. Frank,
Ph.D., and Thomas G. McGuire, Ph.D., presents a discussion of the theoretical economic
rationale underlying carve outs and potential effects on the cost, efficiency, and quality of health
care. The authors note that much of the concern about carve outs is related to quality and that
carve outs almost certainly increase administrative costs and problems in coordinating services.
They also note that carve outs may improve efficiency because the entities providing services
often enjoy significant economies of scale. The authors maintain that the strongest rationale for
carve outs is to increase efficiency.
- "Carve outs: Definition, experience, and choice among candidate conditions," (pp. SP45-57)
by David Blumenthal, M.D., M.P.P., and Melinda Beeuwkes Buntin provides an overview of the
structure, organization, and functioning of carve outs and discusses the diseases and
circumstances for which carve outs are likely to be most advantageous. The authors believe that
carving out services may improve the performance of markets by reducing opportunities for risk
selection. They also argue that carving out specific services may increase the quality of care by
funneling patients to high-quality providers and by increasing the volume of a specific service
offered by a provider. In addition to administrative and coordination problems, the authors note
the potential for problems associated with separating vulnerable populations and problems related
- "Behavioral health services: Carved out and managed," (pp. SP59-67) by Saul Feldman,
D.P.A., notes that behavioral health services traditionally were not covered in the health insurance
benefit package. Today, behavioral health care services are increasingly being covered, but they
are carved out in the sense that employers and plans charged with managing the entire benefit
package contract with specialized for-profit plans (managed behavioral health care organizations
[MBHCOs]) to manage the behavioral health component. Today, MBHCOs oversee the
behavioral health benefits of more than 100 million people, and according to the author, available
evidence indicates that they have been successful in reducing costs between 30 and 40 percent.
He cites empirical evidence to suggest that behavioral health carve outs have increased access and
- "Cancer carve outs: Specialty networks and disease management: A review of their evolution,
effectiveness, and prognosis," (pp. SP71-89) by Bettina Kurowski, Ph.D., examines why the
success of cancer carve outs has been modest, even though cancer might appear to be an ideal
condition for a carve out. The author attributes the limited use of carve outs for cancer care to the
characteristics of the condition and features of health care markets. In particular, she notes that
cancer is not one disease but almost 100 different, complex diseases, making it more difficult to
manage clinically. Also, patient preferences and clinicians' decisions that incorporate them are
likely to vary significantly. And, given the seriousness of the disease, these preferences are hard to
ignore. Thus, treatment guidelines are difficult to establish and follow. Finally, the author
discusses the financial incentives and disincentives of cancer carve outs to plans and specialty
providers. Treatment of a single patient can involve a number of specialties that are difficult to
include in a single-price package.
- "Massachusetts Medicaid and the Community Medical Alliance: A new approach to
contracting and care delivery for Medicaid-eligible populations with AIDS and severe physical
disability," (pp. SP90-98) describes the experience of the Community Medical Alliance (CMA) in
Boston, MA, in treating severely ill Medicaid recipients on a prepaid basis. Currently, CMA treats
204 severely disabled individuals, 141 people with advanced AIDS, and 76 people with general
disabilities or mental retardation. Reimbursement rates for these categories were set by
Massachusetts in 1992, and CMA does not place its providers at financial risk for the services
they order and does not rely heavily on utilization review. CMA focuses on prevention, promotion
of education and self-management strategies, and the provision of timely social support and
mental health services. CMA relies heavily on primary care physicians-who have absolute power
to schedule specialty consultation and admit patients to the hospital—and nurse
practitioners—who have prescriptive authority and ample freedom to order diagnostic and
therapeutic services. Although CMA has operated successfully in a capitated environment, it is
unclear whether CMA's experience is generalizable. In States where Medicaid contracts are
awarded to the low-cost bidder, it is not clear that CMA's model of care is feasible.
Workshop participants called for additional research in five fundamental areas: models of
carve-out organizations and arrangements, objectives of payers and consumers, effects of carve
outs on costs of care, effects of carve outs on quality, and the ways in which carve outs improve
and/or impede access to care. Participants noted several challenges to conducting such research,
including the proprietary nature of data collected and maintained by carve-out organizations, the
need for a comparative conceptual framework, and the need for outcome measures and
For more information, see the June 15, 1998, special issue (volume 4) of the American Journal of
Managed Care for the full text of the papers described here.
Reprints of the overview article (pp. SP11-21) "'Carve-outs' and related models of contracting
for specialty care: Framework and highlights of a workshop," by Drs. Friedman, Devers,
Hellinger, and Fraser (AHCPR Publication No. 98-R080) are available from the AHCPR Publications Clearinghouse.
Return to Contents
Health Care Costs and Financing
More than 4.5 million children lack insurance despite Medicaid
Approximately 4.7 million children are eligible for Medicaid health insurance but are not enrolled
in the program and have no health insurance benefits, according to a recent study by the Agency
for Health Care Policy and Research. In other words, two of every five uninsured children in the
United States could have Medicaid insurance if they or their parents applied for it. Medicaid is the
Federal-State program that provides comprehensive health benefits for low-income people.
The study reveals that 21.2 million children ages 18 or younger were eligible for Medicaid in the
first half of 1996, but 22 percent of them (4.7 million) were not enrolled in the program and had
no other public or private health insurance. This number is far more than previously estimated and
comes after a decade of changes have broadened the criteria by which children qualify for
Medicaid coverage, note AHCPR researchers Thomas M. Selden, Ph.D., Jessica S. Banthin,
Ph.D., and Joel W. Cohen, Ph.D.
Expanded Medicaid eligibility requirements have nearly doubled the number of eligible children,
according to the authors. They note that children who received cash assistance through the old
welfare program—Aid to Families with Dependent Children—were automatically enrolled in
Medicaid. But teenagers and children who became eligible for Medicaid because of recent
program expansions were less likely to be enrolled.
It may be that these families were less aware of their Medicaid eligibility, since they were
ineligible for cash assistance, explain the researchers. In addition, these families may have resided
in neighborhoods with lower Medicaid prevalence, which perhaps increased their sense of stigma
associated with Medicaid. The findings reported in this study are based on analysis of data from
the 1996 Medical Expenditure Panel Survey (MEPS), a stratified random sample of households
designed to yield a nationally representative estimate of insurance coverage, medical expenditures,
and other health-related and socioeconomic characteristics.
Details are in "Medicaid's problem children: Eligible but not enrolled," by Drs. Selden, Banthin,
and Cohen, in the May 1998 Health Affairs 17(3), pp. 192-200.
Reprints (AHCPR Publication No. 98-R067) are available from the AHCPR Publications Clearinghouse.
Organizational strategies, not physician incentives, may be the key to HMO outpatient cost savings
Organizational strategies that influence HMO patients' selection of physicians and provide
physicians with feedback on the cost-effective use of resources, rather than financial incentives to
the physician, may be the key to cost savings achieved by health maintenance organizations
(HMOs), suggest Ann Barry Flood, Ph.D., of Dartmouth Medical School, and her colleagues. In a
study supported in part by the Agency for Health Care Policy and Research (HS06159), they
analyzed health care use and costs for all adults under age 65 being treated for an episode of any
of seven illnesses (100,000 episodes of care) at a large group practice that treated both HMO and
traditional fee-for-service (FFS) patients during the 3-½-year study period (1986 to 1989).
The group practice compensated its physicians based primarily on services rendered without
regard to patient insurance. HMO and FFS patients had access to the same physicians and same
facilities; and at the individual level, providers had identical financial incentives.
The financial incentives aimed at HMO patients required them to pay $10 for each visit to a
physician, which was waived for services provided by midlevel providers or increased to $50 if
services were provided in the emergency room (ER). The HMO used three additional strategies to
encourage patients to see generalists. First, all HMO patients had to select a gatekeeper physician
from a designated pool of providers. Second, although the pool included subspecialists, newly
enrolled patients found that generalists were readily available while subspecialists' practices were
usually full. Third, the practice expanded its branch clinics where patients could access primary
care in more convenient locations. In contrast, specialist care was housed in the less convenient
As intended, HMO patients with the same disease as their FFS counterparts made more use of
lower cost providers (more midlevel providers and generalists), avoided higher cost providers
(fewer specialists and ER visits), and received arguably more cost-effective services (less
expensive ancillary services and more selective use of specialists). For all seven diseases studied,
HMO care had fewer expenditures, adjusted for case mix, with most of the savings occurring for
patients with average health. The healthiest and sickest patients tended to receive the same total
outpatient resources regardless of insurance type.
According to Dr. Flood, these findings suggest that HMOs can achieve significant outpatient
savings by influencing which physicians treat HMO patients.
More details are in "How do HMOs achieve savings? The effectiveness of one organization's
strategies," by Dr. Flood, Allen M. Fremont, M.D., Ph.D., Kinam Jin, Ph.D., and others in the
April 1998 Health Services Research 33(1), pp. 79-99.
Some health care providers respond to reduced Medicare payments by changing the volume of
procedures they perform
For Medicare purposes, over 200 medical procedures were deemed to be overvalued following
implementation of the Omnibus Budget Reconciliation Acts of 1989 and 1990. Medicare reduced
payment to health care providers for these procedures, while payments from private payers
remained relatively high. Eight medical specialties and individual providers within these specialties
varied in their response to lower Medicare payments in terms of volumes of procedures they performed, finds a study supported in part by the Agency for Health Care
Policy and Research (HS08046).
Ming Tai-Seale, Ph.D., of Indiana University, and colleagues analyzed discharge data from about
200 hospitals in the United States over 45 months to determine procedure volume responses from eight specialties experiencing varying
degrees of Medicare payment reductions for selected procedures. They found that the relative
payment margins from payers (amount that reimbursement from the payer was lower than
procedure charges) played a significant role in how physicians reacted to payment reductions from
one payer. Further, the ability of physicians to generate additional demand varied across different
kinds of procedures. Diagnostic procedures and procedures that were less invasive experienced a
greater increase in volume, whereas surgical procedures that were more invasive or more constrained by demand experienced less or no volume response.
Among the eight specialties—cardiology, gastroenterology, general surgery, gynecology,
ophthalmology, orthopedic surgery, thoracic surgery, and urology—gastroenterologists provided
more non-overvalued substitute procedures to privately insured patients. Gynecologists provided
more non-overvalued substitute procedures to Medicare patients and less overvalued procedures to Medicare patients. Urologists provided
more overvalued procedures to both Medicare and privately insured patients, as well as
non-overvalued substitute procedures to Medicare patients. The authors conclude that fixing only
one payer's reimbursement policy is at best a partial solution to cost containment.
Details are in "Volume responses to Medicare payment reductions with multiple payers: A test of
the McGuire-Pauly model," by Dr. Tai-Seale, Thomas Rice, Ph.D., and Sally Stearns, Ph.D., in
Health Economics 7(3), pp. 199-219, 1998.
Return to Contents
Proceed to Next Section