In an effort to achieve cost-effective disease
management, health care systems are developing standards of care and
practice guidelines for oncologists. Although some physicians have
begrudged the inescapable trend, speakers at this ASCO educational
session explained how such guidelines, or care pathways,
for common cancers are positively affecting the quality of cancer care.
NCCN Treatment Guidelines
The National Comprehensive Cancer Network (NCCN) has published
practice guidelines for the treatment of most common cancers. These
have been enthusiastically received. Patient versions of NCCN
guidelines for breast, prostate, and colorectal cancers also have
According to Jane Weeks, MD, who moderated the educational session,
the NCCN has been remarkably successful in ensuring buy-in
among both the clinical and research communities. A key factor in
this success, she said, is that the content of the guidelines strikes
the right balance between strict evidence-based medicine and a
consensus-based clinical approach. Frequent updating and the
inclusion of patient input have also contributed to acceptance of the guidelines.
Clearly, there is interest in this concept. The NCCN
patient-oriented breast cancer guidelines, in fact, received 1
million hits the first day it was online, Dr. Weeks remarked.
NCCN Outcomes Databaes
In addition to developing guidelines, it is also critical for
institutions to produce and utilize their own prospective database,
evaluate adherence to the guidelines, and assess outcomes under the
guidelines. The data from this exercise will help identify the most
effectiveand especially the most cost-effectivestrategies
for managing the common cancers. This is the NCCNs
research mission, Dr. Weeks noted.
The same data, if it is of sufficient quality, can be used for two
different purposes. The first is to evaluate providers, which
addresses quality-of-care questions. The second is to evaluate
treatments and interventions, which answers the research questions,
Dr. Weeks pointed out.
A project was initiated to collect uniform sociodemographic,
clinical, treatment, and outcomes data for women treated for breast
cancer in the NCCN. Breast cancer was selected as the first disease
for study because of its importance as a public health problem, and
because it accounts for a large proportion of overall clinical
activities in the NCCN. Initially, five institutions participated in
this project. In the future, there are plans to include more NCCN
institutions, and other disease sites. We also are very anxious
to bring some community institutions into this project to begin
getting a more heterogeneous data set for research purposes. We want
to see whether these methods can be exported out of the large
tertiary care centers into the real world, where most of oncology is
practiced, Dr. Weeks told the audience.
In addition, the NCCN will seek partnerships with pharmaceutical and
biotechnology industries for financial support. I think we need
to start incorporating standards-of-care into the fabric of what we
do, she suggested. And we cant expect the doctors
or the hospitals to simply pick up the expenses for this out of good will.
M.D. Andersons Mitchell Morris, MD, commented on the value of
the NCCN guidelines. They are good starting points for private
oncologists or institutions that want to begin thinking
systematically about how to treat the most common cancers. They
set a foundation for acceptable practice but allow for variability
and tailoring to individual needs, he said.
Assessing Quality of Care and Performance
Cancer care in the United States is good, but how can it become
absolutely superb? asked Lee Newcomer, MD, of UnitedHealth
Group. He suggested three main ways to improve quality: (1) measure
physicians performance and correct deficiencies; (2) for rare
cancers or complex disease states, use centers of excellence; and (3)
when evidence-based data are lacking, enroll patients in clinical trials.
As we do these things, we will start improving cancer care, and
it will be a lot more fun to practice medicine, he remarked.
Dr. Newcomer approached the topic of performance measurements by
describing a personal experience. Three months after I went to
work for the HMO, my boss called me in and said she would like to
spend 30 minutes reviewing my performance. She went over a number of
objective parameters pertaining to things Id done well, and
things I could continue to improve on. What struck me was that I had
practiced oncology for the past decade without a single performance
review, without seeing any objective performance data, and for that
matter, without even a subjective performance review. I used
the standards that most clinicians in private practice usedwas
my office full, did I have a new set of referrals each day, did I
take care of physicians families?and since I answered
yes to those questions, I assumed I was performing very well.
Dr. Newcomers experience was not unusual, but the assumption
was wrong. According to the UnitedHealth Group database, physicians
comply with published standards of care only about 50% of the time.
Its not because they are particularly bad physicians.
Rather, its because they have never measured their performance.
They may not even realize they have a problem, so obviously they have
not done anything to solve it, explained Dr. Newcomer.
One needs only to go to the medical literature to substantiate this
lack of compliance with standards of care by oncologists. A study by
Timothy Lash in the mid-1980s found that 16% of patients with breast
cancer were not correctly staged, and another 18% did not receive
definitive adjuvant therapy; not unexpectedly, mortality for
inaccurately staged patients was 80% higher than mortality for
correctly staged patients.
The Physicians Data Corporation, which extracts data from transcribed
notes and creates practice profiles, painted a similar picture.
However, taking the procedure out of the hands of the average
physician is not easy, Dr. Newcomer conceded. Nevertheless, I
think an honest discussion about this at your tumor board, medical
staff meeting, or department meeting is called for. Start thinking
about how to handle the unusual or rare cases, he advised the audience.
Participation in Clinical Trials
Finally, when the answer is not obvious or the condition is unusual,
Dr. Newcomer believes that patients should be considered for
participation in clinical trials. The National Randomized
Trials for Breast Cancer announced at ASCO that 1,300 women
participated in trialsbut this means an additional 12,000 women
were treated off-trial, from whom we will never get scientific
information, he noted.
It is not only the quantity of trials that should be encouraged but
also the quality, Dr. Newcomer added, noting that small,
nonrandomized trials and case series compete for the same patients as
national trials and do not produce the caliber of results that can
contribute to clinical decision-making. As a case in point, he said
that the 1997 May/June issue of Cancer listed 29 inadequate
clinical trials involving stages II, III, and IV breast cancer.
At present, the majority of cancer patients initiate enrollment in
clinical trials themselves. Although oncologists agree in
principle that they would like to put patients on trials, it
doesnt happen, Dr. Newcomer observed. And it is not for
lack of coverage. In June 1993, for example, UnitedHealth waived its
exclusion for experimental trials and instituted coverage for any
patient enrolling in a national breast cancer trial. We had 2
women take us up on that offer, while 600 women were treated
off-trial with bone marrow transplants during that time, he
said. In December 1998, UnitedHealth expanded coverage to include any
cooperative cancer center trial, but only eight patients have
enrolled in 6 months, six being self-referrals.
Their oncologists did not recommend trials at all. We need to
get rid of the duplication, to prioritize good trials, and to make
sure that the patients get on them, Dr. Newcomer stated.
The M. D. Anderson Experience
Patients at M. D. Anderson Cancer Center who are not included in
clinical trials are treated under practice guidelines or care
pathways derived from research-driven patient care, better
known as evidence-based medicine. These programs encompass disease
management guidelines and performance measurements, and yield outcome
information that can give clini- cians feedback and provide data for
further institutional efforts at improvement, according to Mitchell
Morris, md, who described M. D. Andersons disease management
system to the audience.
The process evolves as follows: a disease site core group reviews the
medical literature pertaining to the condition, agrees on basic
treatment parameters throughout the continuum of care, and writes
guidelines for each subspecialty component of care. Every clinical
oncologist at M. D. Anderson reviews the content, after which there
is a final peer review process before the guidelines are put in
place. These care paths not only guide the day-to-day disease
management of patients but also allow for the assessment of overall
survival, disease-free survival, toxicity, treatment cost, quality of
life, and other functional assessmentsie, they provide the
all-important outcome measurements.
According to Dr. Morris, the program is basically a means of
formatting and using information that is already being gathered.
There is not a single element in our database that we dont
already collect when a patient comes in with a history and physical
and treatment plan, he said. Its just that we
normally collect this information in a format that is relatively
useless for structured data.
The new system captures and makes more efficient use of structured
data, ultimately eliminates waste (eg, by eliminating tests that do
not contribute to outcome), and encourages adherence to the most
rational treatment plan. This reduces treatment costs in the long run.
To ensure that the guidelines are readily available and useful, they
are provided in an Internet-based system that allows for online
viewing, patient enrollment, and disease management. Utilization is
controlled through physician order sets (sets of physician orders
that outline the tests and procedures [eg, chemotherapy, surgery] for
a plan of care). Using order sets reduces variability in ordering
patterns by eliminating unnecessary tests and treatments. This has
been shown to both lower costs and increase quality of care.
Reductions in Length of Stay, Resource Utilization
Guidelines have been completed for most of the major cancer
diagnoses, and almost 12,000 patients have been enrolled in 70
completed care paths to date. In most disease areas, the program has
reduced length of hospital stay and resource utilization. Surrogate
measures of quality have shown that, even with reductions in costs
and earlier hospital discharges, the quality of care has actually
been maintained or has even improved, Mr. Morris reported.
For example, in an evaluation of hysterectomies for uterine cancer
during three time periods, analysts found that median hospital stay
was shortened from 6 to 4 days after the care path was accepted (and
is now at 3 days). Laboratory costs were reduced by 35%, and total
hospital costs dropped by 20%. There were no readmissions for
complications, and a survey indicated that patients were satisfied
with length of hospital stay and with the medical personnel.
Similarly, an evaluation of colon resections found that average
length of stay was 5.6 days for patients on the care path, compared
to 8.7 days for patients treated otherwise; average total costs were
$12,000 and $20,000, respectively. The program also streamlined care
on the pulmonary service, eliminating 2,000 unnecessary chest x-rays
annually without an increase in mortality.
Although the program was enthusiastically embraced by most
administrators and some early adapter oncologists, not
every physician was immediately sold on the concept, Dr. Morris
admitted. As we started to roll out this program of practice
guidelines and care paths, we ran into, as one might expect,
significant resistance from the physician community. They claimed it
was cookbook medicine, represented a loss of
independence, and would not be applicable to their more
complicated patients. We pointed out, This is not a
rigid system but a framework or a guideline for care....We had
to show the physicians some data to convince them of the benefits of
Dr. Morris cited several key factors as critical to M. D.
Andersons success in creating care pathways for common cancers:
(1) establish a sense of urgency with physicians, explaining the need
for change; (2) make this a physician-driven process, not a
nursing exercise; (3) obtain adequate financial
resources; (4) share a group vision and work in a multidisciplinary
fashion; (5) keep quality of care, and not cost-savings, as the
focus; (6) maintain evidence-based medicine as the programs core.
Future directions at M. D. Anderson may include the alignment of
financial incentives with quality-of-care goals. We have to
provide value for patients, for payors, and for doctors as well, and
I think we are approaching that, he said.