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 It was current when produced and may now be outdated. Archive material is no longer maintained, and some links may not work.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.
Medicare beneficiaries will be among the first to
benefit from AHCPR's CAHPS® initiative
A new nationwide effort is now underway to help patients rate their health plans and to help
consumers choose among plans. The effort is built on a new survey and reporting tool, the
Consumer Assessment of Health Plans (CAHPS®), that provides detailed information on the
experiences consumers have had with their health care providers and plans (select for press release).
The Agency for Health Care Policy and Research supported development of the new CAHPS®
survey, which can be used by managed care plans, employers, and others to obtain consumers'
views of the care they are receiving. Medicare beneficiaries are among the first to benefit from
CAHPS®. The Health Care Financing Administration is using CAHPS® to conduct the first-ever
survey of beneficiaries in every Medicare managed care plan to assess their experiences with
About 130,000 Medicare managed care enrollees have been sent the CAHPS® "Medicare
Satisfaction Survey" questionnaires. Results from the survey will let beneficiaries make
comparisons of managed care plans based on consistent data across plans. They will be able to
find out in a clear, reliable way what other beneficiaries think about the health care plans they
have been using.
Medicare enrollees can choose to receive health care services through traditional fee-for-service
coverage or through managed care plans. After this initial survey of a sample of managed care
enrollees in each of the Medicare-qualified plans, a similar effort will be undertaken with Medicare
beneficiaries in fee-for-service plans.
The CAHPS®-based Medicare questionnaire is to be returned by beneficiaries in May. Results are
expected to be made available by November via the Internet.
In addition, the Office of Personnel Management (OPM) will adopt CAHPS® for use by the
Federal Employees Health Benefits Program. OPM will use CAHPS® to survey Federal employees
and report back the findings of the survey to the employees to help them choose among health
plans during the Federal open season.
The CAHPS® kit contains a set of scientifically based survey instruments that ask about consumers'
experiences with their health plans, sample formats for reporting results to consumers, and
detailed instructions to help implement the surveys and produce the reports. The questionnaires
have been standardized so that assessments can be compared across health plans and patient
populations and over time. CAHPS® also contains survey questions that address health care of
children, people with chronic conditions, and other special populations.
As pointed out by AHCPR Administrator John M. Eisenberg, M.D., some aspects of a plan's
quality can be judged only by the consumers, such as how well a doctor communicates with his or
her patients, how easy it is to get care, or how burdensome the paperwork is.
With AHCPR funding and support, CAHPS® is being developed by a consortium led by Harvard
Medical School, RAND, and the Research Triangle Institute. The 5-year project is in its third
year. All of the CAHPS® products are being evaluated and tested at a number of demonstration
sites to ensure accurate comparisons in real world settings.
To find out more about this new survey and reporting tool, request a CAHPS® Preview Kit
(AHCPR Publication No. 97-0012) from the AHCPR Publications
Clearinghouse. For online information about CAHPS®, select Quality Assessment.
Return to Contents
Managed Care/Market Forces
Managed care plans affect care primarily through
utilization review and discounted fees
Managed care plans provide a variety of incentives and constraints to physicians that influence
their delivery of care. Employers and consumers should be aware of these features before
choosing a health insurance plan, caution researchers who just completed a survey on the topic.
The two most widely used managed care techniques were utilization review, applied to an average
of 59 percent of patients, and discounted fees, applied to an average of 38 percent of patients,
according to the 1995 survey of 2,003 U.S. physicians. This research, led by Dahlia K. Remler,
Ph.D., of Tulane University, was supported in part by the Agency for Health Care Policy and
Research (National Research Service Award training grant T32 HS00020).
An average of 16 percent of patients of doctors surveyed were covered by health insurance
contracts linking physician compensation with use of services. Also, an average of 59 percent of
their patients had their length of hospital stay reviewed, 45 percent had the site of care reviewed,
and 39 percent had the medical appropriateness of treatment reviewed. Among surveyed
surgeons, an average of 49 percent of their patients had site of care reviewed, and 43 percent of
patients had treatment appropriateness reviewed. However, only about 3 percent of the types of
care studied were ultimately denied, with coverage most frequently denied for mental health,
substance abuse, and referral to a specialist of choice.
Use of managed care techniques varied more within States than between States, but conventional
measures of HMO market penetration revealed little about how managed care affects physicians.
For example, California and Massachusetts had nearly the same level of HMO penetration. But
while many California physicians frequently faced strong financial incentives, especially capitation,
to encourage them to provide less care, such techniques were rare in Massachusetts. Thus, the
change from fee-for-service to managed care has very different effects in different places.
For more information, see "What do managed care plans do to affect care? Results from a survey
of physicians," by Dr. Remler, Karen Donelan, Sc.D., Robert J. Blendon, Sc.D., and others, in the
Fall 1997 Inquiry 34, pp. 196-204.
Coverage of new medical technology depends on the
type of insurance plan and the circumstances
It has been assumed that fee-for-service (FFS) insurers will cover any new technology that the
patient and doctor agree to use, perhaps unnecessarily raising the costs of care without necessarily
improving patient welfare, in contrast to health maintenance organizations (HMOs) which are
assumed to be less likely to cover such costs. This assumption is false, conclude researchers
supported by the Agency for Health Care Policy and Research (National Research Service Award
training grant T32 HS00009).
The authors point out that FFS insurers do refuse to cover new technologies they label as
"experimental" or "discretionary." Recent examples include coverage of in vitro fertilizations
(regarded as effective but discretionary), bone marrow transplantation for breast cancer (regarded
as experimental), and coverage for prostate specific antigen (PSA) tests for prostate cancer
(regarded as effective but not cost-effective). HMOs often will agree to cover all the procedures
that patients "need" but then reserve for themselves the power to define what "need" is,
according to the researchers.
The investigators point out that although both FFS and HMO plans can refuse to pay for certain
new technologies, they differ in their approaches. Once an FFS insurer has agreed to cover some
new technology, it must pay benefits for whatever quantities of care the physician and patient
select. In contrast, the HMO plan requires no more than nominal cost sharing, but it can set limits
on the quantities of services it will provide and pay for. These limits may constrain both the
doctor and the patient and may be different for different illnesses.
This depiction of FFS and HMO approaches is based on the researchers' model of two types of
insurance—one (HMO) that controls cost through supply-side quantity limits, and one (FFS)
controls use through demand-side cost sharing. Their model suggests that neither type of
insurance is intrinsically more "inflationary" than the other. Depending on the circumstances,
HMOs may adopt costly technology that FFS insurers do not and vice versa. In addition, even
when both plans approve use of a technology, the level of use approved may sometimes be lower
in FFS plans and at other times lower in HMO plans.
See "Structural incentives and adoption of medical technologies in HMO and fee-for-service
health insurance plans," by Scott D. Ramsey, M.D., Ph.D., and Mark V. Pauly, Ph.D., in the Fall
1997 Inquiry 34, pp. 228-236.
AHCPR report shows businesses can negotiate with
their health plans for both quality and lower costs
Employer market power can be a major force for promoting health care quality and value,
according to a new report by the Agency for Health Care Policy and Research. Theory and
Reality of Value-Based Purchasing: Lessons from the Pioneers—prepared as part of the
research initiative on employers and health care—describes how a number of major
as well as business and health coalitions, are using their market power to promote quality by
collecting data on quality as well as cost, using the data to select health plans and providers, and
offering employees financial incentives to enroll in plans with good performance records. Some of
the organizations are even working directly with providers to identify and implement best
AHCPR-funded research shows that employers in some parts of the country are teaming up with
government purchasers, such as State Medicaid programs and local health departments, to
develop common strategies for encouraging quality care in their communities, according to Irene
Fraser, Ph.D., Director of AHCPR's Center for Organization and Delivery Studies.
The pioneering groups whose activities are described in the report are the Dallas-Fort Worth
Business Group on Health, Chicago Business Group on Health, Gateway Purchasing Association,
General Motors, Digital Equipment Corporation, GTE Corporation, Pacific Business Group on
Health, Iowa's Community Health Purchasing Corporation, and Pacific Bell.
The report also says that most employers continue to focus exclusively on the cost of health plans
and are doing little or nothing to integrate quality with health care contracting. Authors Jack
Meyer, Ph.D., Lise Rybowski, M.B.A., and Rena Eichler, Ph.D., also found that a middle group,
whom they call "dabblers," are taking the first steps toward value-based purchasing by beginning
to collect some quality information from plans and providers.
Dr. Meyer, the lead author, is president of New Directions for Policy, a Washington, DC-based
health care consulting firm; Lise S. Rybowski, M.B.A., is founder and principal of Severyn
Healthcare Consulting, a health care research and communications firm in Fairfax, VA; and Rena
Eichler, Ph.D., is a health economist with AHCPR's Center for Organization and Delivery
Print copies of Theory and Reality of Value-Based Purchasing: Lessons from the Pioneers (AHCPR
Publication No. 98-0004) are available from the AHCPR Publications
A related report, Public-Private Healthcare Purchasing Partnerships, which resulted from an
AHCPR-sponsored conference, is available from the Midwest Business Group on Health, 8765
West Higgins Road, Suite 280, Chicago, IL 60631; phone (773) 380-9090.
Return to Contents
Making residents more aware of the costs for lab tests
and x-rays does not change their test-ordering
Lowering the costs of care is a major concern for all sectors of the health care industry, and
diagnostic testing represents a significant proportion of hospital costs. A variety of studies
performed over the past 20 years have found that many of the clinical laboratory tests and
radiologic procedures performed in teaching hospitals are unnecessary.
A recent study supported by the Agency for Health Care Policy and Research (HS08297)
examined whether a computerized display of charges at the time of test ordering would affect the
number of tests ordered by doctors for hospitalized patients. David W. Bates, M.D., M.Sc., and
his colleagues at Brigham and Women's Hospital performed two prospective controlled trials.
Each trial included all medical and surgical inpatients at a large teaching hospital: 3,536
intervention and 3,554 control inpatients over a 4-month period made up the group with clinical
laboratory tests, and 8,728 intervention and 8,653 control inpatients over a 7-month period
formed the group with radiological tests.
Patients whose clinical laboratory test charges were displayed had 4.5 percent fewer tests ordered
and 4.2 percent lower total charges for tests than the control group, but the differences were not
statistically significant. The number of radiological tests ordered and the charges for those tests
were almost identical for patients in the intervention and control groups.
These findings for hospitalized patients contrast with those from previous studies of diagnostic
tests ordered for outpatients in which the display of charges significantly reduced the number of
tests ordered. It may be that hospitalized patients are sicker than outpatients, and thus tests are
less likely to be discretionary. On the other hand, the physicians in this study were residents who
had no financial stake in the decision. Thus, they might be expected to be affected less than a
physician group with direct financial incentives to reduce use.
Physicians liked seeing the charge figures, and the nonsignificant decrease in the intervention
group, if annualized, would result in $1.7 million lower laboratory charges for the hospital. Thus,
the hospital has continued to display this information.
The researchers conclude that to decrease the ordering of ancillary tests substantially (more than
10 percent), stronger and more intrusive interventions may be necessary, for instance the
suggestion of a less expensive substitute at the time of order or problem-oriented, test-ordering
guidance for clinical problems.
See "Does the computerized display of charges affect inpatient ancillary test utilization?" by Dr.
Bates, Gilad J. Kuperman, M.D., Ph.D., Ashish Jha, M.D., and others, in the November 24, 1997
Archives of Internal Medicine 157, pp. 2501-2508.
Hospital complication rates and mortality rates may
not correlate when severity of illness is taken into
Cost-driven hospital restructuring initiatives in the United States and other developed countries,
which typically involve reducing nursing staff, have been criticized for their potential to diminish
quality of care. A hospital's mortality rate for a given condition is often used as an indicator of its
quality of care, but patient deaths may not be numerous enough to have the statistical power to
distinguish quality of care among different hospitals. Nevertheless, substituting complication rates
for mortality rates to measure hospital quality of care, as some have suggested, would be a
mistake, according to University of Pennsylvania researchers Jeffrey H. Silber, M.D., Ph.D., and
Paul R. Rosenbaum, Ph.D. Their study, which was supported by the Agency for Health Care
Policy and Research (HS06560), shows that complication rates often do not correlate with
The researchers studied both mortality rates and complication rates using three predictive models
that varied in the degree to which adjustments were made for patient severity of illness. They
tested the models on individuals registered in a 1991 national database with gastrointestinal tract
diseases, liver diseases, and those who underwent breast biopsy and mastectomy procedures. Drs.
Silber and Rosenbaum found that hospital complication rates and mortality rates were well
correlated when not adjusted for patients' severity of illness. However, when clinical data on
severity of illness were added to the models, the correlation disappeared.
The researchers conclude that inadequate control of variables such as patient severity of illness
may lead to a spurious high correlation between an accepted outcome measure (mortality rates)
and a new outcome measure (complication rates). Their findings were published in a paper
presented at the Bellagio Conference on Hospital Restructuring in North America and Western
Europe, held in November 1996, which was convened to determine the extent and impact of
hospital restructuring in the U.S., Canada, England, Scotland, Germany, and The Netherlands.
The conference papers and a foreword by Carolyn M. Clancy, M.D., Director of AHCPR's
Center for Outcomes and Effectiveness Research, and Marcy L. Gross, Senior Advisor on Women's Health, are presented in a recent journal supplement published with AHCPR
More details of this paper, "A spurious correlation between hospital mortality and complication
rates: The importance of severity adjustment," by Drs. Silber and Rosenbaum, as well as the
foreword and other conference papers, may be found in the October 1997 Medical Care
Supplement 35(10), pp. 0S77-0S92.
Level of threat determines future course of rural
Rural hospitals that are moderately or highly threatened with recurrent factors faced by hospitals
and their leaders are more apt to convert to another type of care, such as acute psychiatric care,
than other, less-threatened hospitals. Or they may become a primary care center or nursing home;
some will close. Those with only slight threats are more apt to specialize in a specific health area,
according to a study supported in part by the Agency for Health Care Policy and Research
Rita Harmata, Ph.D., and Richard J. Bogue, Ph.D., of the American Hospital Association,
examined 16 cases of hospital change: 5 specializations, 6 conversions, and 5 closures. This
included telephone interviews with each of the 49 key informants who experienced the hospital's
change firsthand. Four recurrent factors faced by hospitals—physician recruitment and
action by local decisionmakers, State and Federal policy, and the local economy—were
used as the
basis for the low-, moderate-, and high-threat categories.
For hospitals that were part of the low-threat category, none of the four factors was pointed to as
a major reason for the change that occurred. In the moderate-threat category, one of these four
factors was singled out, while the high-threat
category included hospitals where those interviewed pointed to more than one of these factors as
a significant reason for the hospital change. In those hospitals studied, conversions and closures
were more typical of the hospitals that had been placed in the moderate- or high-threat categories,
while specializations were more typical of hospitals that were part of the low- or moderate-threat
categories. Four cases fell into the low-threat category, where none of the factors was seen as the
specific reason for the change. In those cases, hospitals were only mildly threatened by events
such as physician dissatisfaction, turmoil among local decisionmakers, and existing Medicare and
Medicaid policies. Three hospitals responded with specialization by attaching intermediate or
long-term care facilities to the hospital, and one closed.
Hospitals in the moderate-threat category all experienced one major threat. In one case, the threat
factor was the local economy. The hospital purchased another local hospital and converted into
outpatient and long-term care services. For another, the threat was a physician crisis which
prompted investment in recruitment and attachment of an 80-bed long-term care facility. When
changes in transportation and trade patterns shifted commerce out of one hospital's area to
hospitals 20 to 25 miles away, it retained its long-term care facility and converted acute beds to
15 independent living apartments and a rural health clinic. Eventually, all six hospitals in the
moderate-threat category had some capacity devoted to nonacute services.
Three conversions and three closures made up the high-threat group. In each case, the local
economy was weakening or too small to support a hospital. One hospital had an increased load of
Medicaid and uncompensated cases in the 1970s and 1980s and finally was forced to close.
Another was doomed by economic consequences of the domestic oil bust in the mid-1980s.
Hospital patients began shifting toward Medicaid and indigent care, recruitment efforts became
expensive and fruitless, and the hospital finally closed. For these hospitals, reductions in
population density, increases in unemployment, and economic downturn were perceived as
determined largely by forces outside the control of the hospital's leadership.
For more information, see "Conditions affecting rural hospital specialization, conversion, and
closure: A case-based analysis of threat and change," by Drs. Harmata and Bogue, in the Spring
1997 issue of The Journal of Rural Health 13(2), pp. 152-163.
Return to Contents
Insulin therapy often does not achieve near-normal
blood sugar levels in patients with type 2 diabetes
For most people with type 2 or adult-onset diabetes, daily insulin injections will not achieve
near-normal blood sugar levels, even in conjunction with additional physician visits and required
home glucose monitoring, concludes a study supported in part by the Agency for Health Care
Policy and Research (HS06665). Type 2 diabetes patients typically are started on insulin if diet,
exercise, and oral medication fail to adequately control their glucose levels. Type 1
(insulin-dependent) diabetics take daily insulin injections to reduce their excessively high blood
sugar levels, which are the hallmark of both types of diabetes. These very high blood sugar levels
can lead to microvascular complications resulting in serious eye, nerve, and kidney disease.
It has been suggested that earlier conversion from oral medication to insulin might achieve better
blood sugar control for patients with type 2 diabetes as well. Patients who have poor control to
begin with apparently do achieve moderate control, which itself greatly reduces diabetic
complications. However, encouraging primary care physicians to increase the use of insulin
therapy seems a suboptimal strategy for decreasing blood sugar further for patients who currently
have moderate glycemic control, notes the Type II Diabetes Patient Outcomes Research Team
(PORT) led by Sheldon Greenfield, M.D., of New England Medical Center. The researchers
merged clinical, survey, and administrative data to examine medical resource use and glycemic
control in a group of 8,668 type 2 diabetes patients cared for from 1990 to 1993 by generalist
physicians in a large health maintenance organization.
Among patients starting insulin therapy, hemoglobin A1c
(HbA1c—the optimal measure
sugar level) decreased by 0.9 percentage point at 1 year compared with those receiving oral
medications. Yet 2 years after starting insulin therapy, 60 percent of patients still had
levels of 8 percent or greater (near-normal levels are less than 7 percent). But patients with a
baseline level of 13 percent had a three-fold greater decline in HbA1c than those
HbA1c level was 9 percent (8 to 10 percent is often termed "moderate" control).
with type 2 diabetes who have the poorest control have the most to gain from insulin therapy.
Patients in this study who were taking insulin had more laboratory tests performed, 2.4 more
outpatient visits per year, and almost 300 more fingersticks for home glucose testing per year
compared with patients using oral agents. The high costs represent a factor to be considered in
taking further action to reduce blood sugar levels.
Select for press release. For the original article, see "Starting insulin therapy in patients with type 2 diabetes," by Rodney A. Hayward, M.D.,
Willard G. Manning, Ph.D., Sherrie H. Kaplan, Ph.D., M.P.H., Edward H. Wagner, M.D., and
Dr. Greenfield, in the November 26, 1997 Journal of the American Medical Association 278(20),
Several office readings of blood pressure may be
needed for accuracy
A one-time blood pressure (BP) measurement in your doctor's office is not sufficient to gauge
your usual blood pressure. To obtain an accurate measurement, readings from at least three office
visits should be averaged, but even this figure can be inaccurate. Even with repeated readings,
office BP level should be regarded as an imperfect estimate of one cardiovascular risk factor,
concludes a study supported by the Agency for Health Care Policy and Research (HS07662).
Normal systolic BP (excluding exercise) can vary by 30-60 mm Hg, and normal diastolic BP can
vary from 20-40 mm Hg over a 24-hour period. Also, some patients react with elevated BP when
the BP reading is taken in the medical setting, the so-called white-coat response, explains Kevin
A. Pearce, M.D., M.P.H., of Wake Forest University.
Dr. Pearce and his colleagues observed 75 randomly selected family practice patients with normal
or high BP to determine the number of routine office visits required to optimize the estimation of
usual (mean) BP in older adults in primary care practice. They defined a person's usual systolic
and diastolic BP levels as the mean awake or mean 24-hour BP calculated from two 24-hour
periods with use of an ambulatory BP monitor; this was used as the standard against which to
compare office BP readings.
Most (76 percent) of the gain in correlation with ambulatory BP in systolic office BP readings was
made at the second visit, with small incremental gains observed through the fifth visit (from .68 at
the first visit to .81 at the second visit to .82, .84, .85, and .85 at later visits. For diastolic BP,
most (80 percent) of the increase in correlation was achieved by the third visit, with small
incremental gains continuing through the sixth visit (from .60 on the first visit to .72 on the third
visit to .73, .74, and .75 by the sixth visit).
Even when 12 office BP readings taken over 6 visits were averaged, in 19 percent of patients
there was a discrepancy of at least 10 mm Hg between estimated and observed usual systolic BP,
despite high correlations between mean office BP and mean ambulatory BP. Office readings were
lower than ambulatory readings almost as commonly as vice versa. The researchers recommend
that clinicians use office BP readings averaged over 3 to 6 visits to estimate the usual BP level of
most patients. The closer the readings are to a decision threshold (to recommend medication or
not, for example) the more readings should be averaged.
See "Comparisons of ambulatory blood pressure monitoring and repeated office measurements in
primary care," by Dr. Pearce, Gregory W. Evans, M.A., John Summerson, M.S., and others, in
the November 1997 Journal of Family Practice 45(5), pp. 426-433.
Patient-physician partnership is the key to better
management of osteoarthritis and other chronic
After age 65, arthritis is the foremost cause of disability. The key to good management of
osteoarthritis—the principal form of arthritis—is for physicians to create an
with patients that helps them to manage their disease, assert Halsted R. Holman, M.D., and Kate
R. Lorig, Dr.P.H., of Stanford University School of Medicine, in a recent commentary. For the
patient, osteoarthritis means far more than pain and limited motion. It causes reduced physical
capability, compromised work capacity, restricted social activities, and heightened emotional
distress. These circumstances are intensified when patients have another chronic disease as well.
Physicians and other health care providers should help patients (and their families) use
medications properly, change behavior to improve symptoms or slow disease progression,
interpret and report symptoms accurately, adjust to new social and economic circumstances, cope
with emotional consequences, and participate in treatment decisions. They also should support
patient efforts to maintain normal activities and adapt new healthy behaviors, recommend the
authors, whose work is supported in part by the Agency for Health Care Policy and Research
In their earlier studies, the authors showed that participation by patients with osteoarthritis in an
arthritis self-management program led to substantial declines in pain and depression, as well as a reduction in the use of health services. After 4 years, patients continued to report substantially
reduced pain and use of health services despite a 9 percent increase in measured physical
disability. The success of this approach appears to be linked to patients' perceived ability to cope
with the consequences of their arthritis as a result of training. A physician-patient partnership that
supports patient self-management should result in patients with fewer symptoms, less emotional
distress, increased activity, and more independent living: what the authors refer to as "successful
For more information, see "Overcoming barriers to successful aging: Self-management of
osteoarthritis," by Drs. Holman and Lorig, in the October 1997 Western Journal of
Medicine 167(4), pp. 265-268.
Return to Contents
Proceed to Next Section