Introduction
and Objectives
A hemodialysis patient’s vascular access is truly his/her lifeline without which hemodialysis would not be possible. Although no type of access in current use consistently provides a blood flow that is adequate to meet any prescription, lasts for the life of the patient and has a low rate of complications, the AVF comes closest to being “ideal” over both AV grafts and indwelling catheters. In 1997, the National Kidney Foundation published the Dialysis Outcomes Quality Initiatives (NKF-DOQI) Clinical Practice Guidelines for Vascular Access, which concluded that the AVF is the primary choice of access.
The NKF-DOQI Practice Guidelines workgroup
recommended an AVF placement goal of 50% in incident patients with an AVF
prevalence rate of 40% (Guideline #29).
Hemodialysis vascular access utilization characteristics in U.S.
dialysis population per the 2002 Clinical Performance Measures (CPM) Project
determined that 31% of prevalent patients were dialyzed using an AVF
(October-December 2001 data). Northwest
Renal Network had 45% AVF utilization from that same time period in prevalent
in-center hemodialysis patients. This
data was validated by the more complete, facility-specific data derived from
the 2001 CDC survey, which reported a 44% prevalent AVF rate in our Network’s
prevalent hemodialysis patients.
Overall our Network’s rates met the K/DOQI >40% AVF
rate guideline for prevalent patients, but the
Northwest Renal Network’s Medical Review Board (MRB) began looking at the issue
more closely as substantial
variation was noted facility to facility.
An opportunity for improvement existed in the range,
which went from a high of 90% to a low of 20% AVF utilization rates. The MRB believed that intervention at these
<40% AVF facilities could improve patient outcomes by not only increasing the
use of AV fistulas, but also reduce the high catheter use rate and associated
complications, including infections, hospitalization and mortality.
After
the opportunity for improving AVF rates was identified, a root cause analysis
conducted by our MRB AVF Sub-committee in February 2002 produced three dominant
barriers to a higher AVF rate. The top
three root causes identified were:
1) Failure of nephrologists to act
as Vascular Access Team Coordinators. This includes making recommendations to
vascular surgeons, assisting in vein preservation and mapping, and working
closely with HD unit staff to assure knowledge and skills regarding access and
cannulation.
2) Lack of AVF
training for vascular access surgeons,
including assessment skills, vein mapping and surgical techniques. Surgical training programs place little
emphasis on vascular access approaches, techniques and troubleshooting,
especially for the more complicated procedures.
3) Late referral, or non-referral,
to nephrology, causing the need for temporary
catheters, which may impact on future access and cause a sense of urgency for a
working permanent access.
The goals for our intervention focused on the first
two root causes. The third root cause
may be indirectly affected by our intervention, but was not a main barrier to
increasing the use of AVF that we concentrated on in this project, nor is this
patient/provider population easily accessible to Network organizations.
The broad goal of this project was to decrease morbidity and improve the quality of life for hemodialysis patients. Since AVFs have been shown to have a longer life with fewer maintenance procedures required than AV grafts and catheters, patients utilizing this form of access require fewer interventions with their concomitant risks. If patients are getting adequate blood flow from their access, having fewer infections, not needing to schedule interventional procedures for complications, then health and quality of life will be improved. There is also financial benefit to the Medicare program as patients experience fewer access-related hospitalizations, do not require additional procedures and medications, and may be more likely to maintain a regular employment schedule. (See Reference list, appendix A)
Northwest Renal Network’s
objective was to raise the AVF rate in our targeted intervention population
(prevalent in-center hemodialysis patients at facilities with less than 40% of
their patients with viable AVFs) by at least two and one half percentage
points. An increase from thirty-one and
one half percent AVFs to thirty-four percent AVFs would meet this goal. We anticipated that this would require at
least seventy-five additional AVFs at the end of 2002 as compared to 2001. We initiated the project with 46
intervention facilities (43% of our facilities) that met the <40% AVF rate
criteria (serving 2,992 patients).
During the course of this project, one of the intervention units closed.
Another important intervention
objective was to convince nephrologists and surgeons to place and maintain more
AVFs in all patients. We recognized
that it would be difficult to observe and measure a large effect in incident
patients during the short time frame of this project. However, we plan to continue to monitor and track prevalent AVF
rates for several years, which should provide data that captures both prevalent
and incident patients in treatment at year end. Prevalent rates will also reflect whether a third intervention
objective was met: to motivate the access team to convert catheters and AV
grafts to AVFs. While we expected some
new patients to receive AVFs and some catheters to be converted to AVFs during
this project, we assumed that the most predictable portion of our impact
(increasing AVFs) will be in conversions from AV grafts to AVFs.
Methods
We targeted
our initial outreach to the 46 facilities in our Network with <40% of
hemodialysis patients utilizing an AVF as of December 2001 (prevalent AVF
rates) per the 2001 CDC survey. (This represented 2992 hemodialysis patients
and 47% of all hemodialysis patients in treatment at all Network facilities as
of December 2001) Our starting point
was to identify personnel and affiliated providers to invite to attend our
educational workshops entitled “Back to the Basics: Increasing the Use of AV Fistulas in
Hemodialysis Patients”.
In February 2002, a letter from our Medical Review Board Chairperson was
sent to the Administrators and Medical Directors of the identified facilities
describing the opportunity for improvement in our Network, the objectives, and
dates and locations of our four intervention meetings. We enclosed a colorful bar graph depicting
all Network #16 facility (blinded) AVF rates, with each respective facility’s
rate highlighted to emphasize their relative standing. We asked these contacts to identify one or
more nephrologists from their staff who could or would lead a vascular access
team and also to identify any vascular access surgeons and interventional
radiologists affiliated with their facility/practice who they thought would benefit
from, or be interested in, the information discussed at the meetings. They sent us names and contact information
on a form we provided for them. In
March 2002, we sent invitations to all whose names we had received from the
above contacts, asking the invitees to submit names of other physicians they
would like to have attend the meeting(s).
From these names we built our database of invitees and attendees.
The
interventions were primarily aimed at physicians. However, to develop the team concept, we asked each intervention
facility Administrator and/or Medical Director to identify a Vascular Access
Manager (VAM), in most cases a nurse, interested in being a key member on a
vascular access team. These identified
VAMs were invited and urged to attend our educational program. We attempted to restrict the number of VAMs
to one per intervention facility, however additional nurses were invited on the
special request of the Medical Directors or Administrators of the intervention
facilities and private physician group practices. These special case VAMs included Directors of Nursing, case
managers and QI nurses from large providers who managed one or more of our
intervention facilities, case managers from managed care organizations who had
patients at these facilities or strong connections with surgeons affiliated
with the intervention facilities, and two office nurses who acted as VAMs with
invited physician groups.
It was
recognized that there would be physicians who treated patients at two or more
facilities, and that not necessarily all of their patients would be served by
our intervention facilities. We tried
to make sure that each intervention facility had at least one representative at
the meeting. All facilities had RSVP’d
that someone would be attending; however, in the end, two facilities did not
have representation at our meetings. The two
units who were not represented were sent the intervention materials and they
were invited to subsequent “non-intervention” meetings, but no other outreach
was made.
The
Intervention Workshop Meetings
As we have learned from much of the
preventive medicine literature, multifaceted or multi-component interventions
have shown the greatest success with physicians. Although we recognized the concerns about the efficacy of
workshops, there was consensus among our Medical Review Board (MRB) and AVF QIP
Subcommittee members that the most cost-effective approach to an intervention,
which targets physicians, would be a time-limited, intensive, peer-to-peer
educational program. We designed a
half-day workshop using a “”Best Demonstrated Practices” (BDP) example from our
own community, sharing the positive results of inter-disciplinary
collaboration. The simple model of
nephrologist/surgeon collaboration and the critical importance of advance
planning proved to be concepts that could be well presented in a focused
workshop. Our predominant meeting
features were physician-to-physician presentations, designed to foster the
development of better working relationships between the key players:
nephrologists and vascular surgeons/interventional radiologists. We chose our meeting locations based on
their proximity to the identified attendees.
We met in Olympia, Washington; Boise, Idaho; Portland, Oregon and
Spokane, Washington.
We provided a Best Demonstrated
Practices example approach to vascular access management, developed in 1996,
tested and proven to work by Dr. Vo Nguyen, a nephrologist from Olympia,
Washington. Dr. Nguyen and his vascular
access surgeon partners, represented by Dr. Chris Griffith at three out of four
of our meetings (his partner Jim Reus, MD presented at one of our meetings),
also of Olympia, have been able to achieve a >90% fistula rate despite all
the usual co-morbid factors dialysis patients present. As described in an abstract Dr. Nguyen
presented at the 2000 ASN meeting, his patient population had an average age of
62.2, was 65% female, 73.5% diabetic, 63% had peripheral vascular disease, 28%
were obese, and 22% were referred late for access placement. He and others have concluded that even in
patients whose initial access is a graft, as that graft fails and a new access
is required, an AVF should be considered.
Patients who come to dialysis urgently (i.e. there has been no time to
place an AVF and allow it to mature) are sometimes not considered candidates
for AVFs. Dr. Nguyen believes this
should not be a deterrent. As noted
above, Dr. Nguyen achieved high AVF rates despite 22% of his patients being
referred late.
Dr. Nguyen attributes his accomplishment
to teamwork, dedication of well-trained and creative surgeons, pre-op mapping
and long-term vascular access planning.
After being encouraged by Dr. Griffith and surgical partners to come up
with a system for planning adequate and effective AV accesses in his ESRD
patients, Dr. Nguyen started from the ground-up re-educating himself. He attended meetings with European peers who
had achieved high AVF rates, observed access surgery at Dr. Griffith’s invitation
and went back to his anatomy books to refresh his assessment skills. He took off with his new skills by
developing a plan for each of his patient’s vascular access transition away
from catheters and AV grafts. Dr.
Nguyen set increasingly ambitious goals for himself each year, and has achieved
them with the help of his surgeon colleagues and the techniques he shared in
his presentation at our workshops. One
of the major strategies that Dr. Nguyen and his colleagues used was conversion
of all AV grafts to AVFs, as the grafts showed signs of deteriorating. His approach is replicable in other practice
settings, and his charismatic presentation was pre-tested and approved by our
Medical Review Board.
Dr. Nguyen, our key speaker,
encouraged nephrologists to take on the role of a vascular access team leader
and recommended that each facility appoint a vascular access manager (VAM) to
oversee a Vascular Access Management Program (VAMP) in their dialysis
units. We have received informal
feedback via comments and formal feedback via our follow-up from the workshop
participants that many were inspired by Dr. Nguyen’s humble and informal, yet
passionate, discussion of his own experience.
Another concept that Dr. Nguyen stresses is the importance of the nephrologist
working closely with his hemodialysis staff to create cannulation experts, or
teaching the patient to cannulate themselves, as a key in maintaining and
preserving a high rate of AV Fistulas.
Dr. Chris Griffith discussed the
current processes of arterial-venous access surgery including the upper arm AVF
transposition technique, Gracz procedure and conversion of AV grafts to
AVFs. He highlighted other solutions to
challenges seen in his own patient population using case studies and
photographs to illustrate his talk. He
also explained his use of Doppler Duplex Scanning in assessing vessels prior to
surgery, sharing his own protocol and mapping diagram.
Carolyn Barclay, RN, CNN, Vascular
Access Coordinator at Optimal Renal Care, LLC, provided information and tools
to help Vascular Access Managers follow and educate patients about vascular
access and the problems that arise in their care, both pre and post ESRD.
Steve Alford, a nurse educator from
Medisystems, instructed the audience on successful cannulation techniques for
working with new AV fistulas, emphasizing the buttonhole technique, which Dr.
Nguyen strongly advocates for maintaining high AVF rates and introducing the
new dull needle for single-site cannulation.
The meeting evaluations were
reviewed after each workshop and comments and suggestions from the attendees,
Network staff and the speakers themselves were incorporated into subsequent
presentations. The speakers were sent
copies of the evaluations and the QI Manager worked with them to integrate
minor changes in the curriculum from one session to the next.
The Network absorbed all the meeting
costs. We did not charge a registration
fee for this meeting and provided continuing education credits (both CMEs and
CEUs) to the participants as an incentive to attend the meeting. We also provided lunch prior to the half-day
meeting and an offer to have “dinner with the speakers” to all attendees
following the workshops to encourage continuation of the discussions in a less
formal venue.
The “take-away” materials we
provided at the meetings were designed to reinforce the key concepts inherent
in effective planning for, and placement and maintenance of AVFs in
hemodialysis facilities and training units.
Many of these tools and materials have also been posted on our Network
website, which can be used as is or modified for individual practitioners’ use,
and/or dissemination to our broader community. (See Materials, attachment B)
The enthusiasm that was generated
by our programs at this meeting, demonstrated by the response of our attendees
and our two main speakers, was so great and the benefit so clear that we
expanded to other regional meetings to invite all of our Network
facility-affiliated nephrologists and surgeons/interventional radiologists to
gain from this information. Our Board
of Directors approved funding for four additional Back to the Basics
meetings, three which took place in 2002, in Anchorage, Alaska; in
co-sponsorship with Northwest Kidney Centers in Seattle, Washington and in
Olympia, Washington. Another meeting is
planned for Lewistown, Montana in May 2003.
Vo Nguyen also made presentations at several ANNA meetings in our
Network states and at meetings sponsored by Renal Care Group.
Quality Outcome Indicators and Results
Our project included two outcome measures:
We
obtained our baseline prevalent AVF outcome data from the facility-specific
2001 National Surveillance of Dialysis-Associated Diseases (CDC 2001) survey,
questions #6 (number of patients dialyzing in the unit in December 2000) and
question #8 (number of prevalent AVFs in the patient population). Historically, CPM surveys have been used to
validate the accuracy of the CDC survey data.
This provided us with data that allowed us to rank our facilities and
identify those with AVF rates <40% which became our intervention facilities,
called “target population” in the following calculations. We collected similar data from questions #2
and #6 from the CDC Survey 2002 and compared it to our baseline data to
determine improvement using the following formulas:
1)
Pre: Numerator:
Number of AVFs in the target population
924
Denominator: Number
of patients in the targetpopulation 2940 31.4%
We removed one facility from the target/intervention group
(10 patients) due to the unit closure.
Post: Numerator:
Number of AVFs in the target population
1196
Denominator:
Number of patients in the target population 2992 40%
During the course of
this project, one facility closed. The numbers above reflect removal of this
facility’s patients from both pre and post measures .
Our original goal was
to achieve at least a 2.5 percentage point increase in the prevalent AVF rate
among hemodialysis patients in our target population (hemodialysis patients
receiving care at our intervention facilities in December 2002). We recognized that we would only have six to
eight months post intervention for change to be reflected and intended to
continue to monitor AVF rates after submission of this report, at twelve months
and eighteen months. Data obtained from the 2002 CDC
Survey reveals a statistically significant 8.6 percentage point increase in the
prevalent AVF rate in our target population (p<0.001, chi-square). (See graph below and attachment C).
We are delighted to see this large a
change this quickly. We suspect that the emphasis placed in our intervention
programs on planning ahead to convert existing grafts to fistulas was a message
heard by attendees and is reflected in this data.
2) Pre: Numerator: Number of AVFs in the comparison population
Denominator: Number of patients in the
comparison
population
Post: Numerator:
Number of AVFs in the comparison population
Denominator: Number of patients in the
comparison
population
Our project used a pre/post with comparison design. Our original plan was to draw baseline and
re-measurement AVF rates for a similar group of facilities (with <40% AVF
rates per 2001 CDC Survey) from three of our neighboring Networks (#15, 17 and
18), in order to measure changes in our target facilities relative to change in
the other Networks’ (AVF <40%) facilities that did not receive
interventions. Several facilities that
were included in Network #15’s AVF intervention project were eliminated. As of writing this report, CDC data from the
other Networks has been delayed due to late release and initiation of the CDC
survey. We plan to compile this data as
soon as it is available and will report it in a follow-up document.
Our project included the following
three Process Indicators
1) Do you have a vascular access manager (VAM) at your facility?
The data for the following
indicators came from the VAM/VAMP Information Requests mailed to all of
our Network facilities with the 2001 and 2002 CDC Surveys in January 2002 and
2003 respectively. Each facility responded “yes” or “no” to whether they had a
staff member who acted as a vascular access manager, and defined the role of
their VAM based on a list of tasks provided on the request. (See attachment
F). Based upon input from our Regional
Office during the planning phase of this project, we established that a minimum
set of three key activities would need to be included in the task set, for this
individual to meet our criteria as truly functioning as an access manager.
Those criteria included: “A” monitoring vascular access in the facility; “B”
acting as the main contact in the unit for vascular access activities; and “C”
reporting vascular access data at QI/QA meetings. Criteria “C” was the most
difficult to achieve In retrospect, we
believe there were flaws in our information-gathering tool which could have
produced unreliable information for this question. Therefore, we are
reporting both the number of facilities
stating that they did have a VAM (regardless of that person’s role), and
the number reporting a VAM whose role met
all of our minimum criteria, and just criteria A & B.. All measures are based upon the number of
facilities responding to our pre and post intervention queries.
Number of Network HD facilities with VAM (“Yes”)Pre-meeting: 27/94 (29%)
Total number of
Network HD facilities responding Post
meeting: 43/106 (41%)
Change: ns
(chi-square)
Number of
intervention facilities with VAM (“Yes”) Pre-meeting: 12/41(29%)
Total number of
intervention units responding Post
meeting: 21/42 (50%)
Change: ns (chi-square)
Number of Network HD facilities with VAM (met
criteria A,B&C) Pre-meeting: 8/94 (8.5%)
Total number of
Network HD facilities responding Post meeting:13/106 (12%)
Change: ns
(chi-square)
Number of
intervention facilities with VAM (met criteria A, B &C) Pre-meeting:
3/41 (7%)
Total number of
intervention units responding Post
meeting:10/42 (24%)
Change:
p<0.05 (chi-square)
Number of Network HD facilities with VAM (met
criteria A & B) Pre-meeting: 18/94 (19%)
Total number of
Network HD facilities responding Post
meeting: 18/106 (17%)
Change: ns (chi-square)
Number of
intervention facilities with VAM (met criteria A & B) Pre-meeting:
6/41 (15%)
Total number of
intervention units responding Post
meeting:13/42 (31%)
Change: ns
(chi-square)
Do you have a vascular access management program (VAMP) at
your facility?
Number of
Network HD facilities with VAMP Pre-meeting: 51/94 (54%)
Total number of Network HD
facilities responding Post meeting: 76/106 (71%)
Change:
p<0.02 (chi-square)
Number of intervention facilities with VAMP Pre-meeting: 22/41 (54%)
Total
number of intervention units Post
meeting: 32/42 (76%)
Change:
p<0.05 (chi-square)
2)
Attendance
at the meetings
We acquired the next four data points from our invitee database and our meeting attendance rosters:
44
intervention units were represented at our intervention meetings = 96%
46
intervention units were invited
Attendance by at least one
representative from each intervention facility was our goal. Staff and physicians at a number of our
intervention facilities had overlapping responsibilities. Thus, one attendee could be counted as
representing several facilities.
Attendance at the workshops was tracked by specialty area and affiliation,
so we were able to note whether any intervention unit had “no presence” at our
workshops.
36
nephrologists attended the four meetings =
37.5%
96
nephrologists were invited to the four meetings
16 VA surgeons attended the four meetings = 17%
93 VA
surgeons were invited to the four meetings
1
interventional radiologist attended one of the meetings =
4%
25
interventional radiologists were invited to the four meetings
Our project included additional self-reported data:
1) Pre meeting/post meeting query regarding
beliefs about AVF:
At each session we
distributed to and nephrologists and surgeons completed pre-meeting queries
to the physician attendees as they registered and post-meeting queries
to them at the end of the meeting. We
used this tool in an effort to determine whether physician’s AVF practice
attitudes changed as a result of hearing our speaker discuss their experience
and results. The attendees returned them
to us on a volunteer basis and we did not track names on these queries, so we
were unable to compare each individual’s attitude change pre and post. We were, however, able to evaluate the
group’s general attitude change pre and post meeting. One of the prominent attitude changes we noted is illustrated
below in our first question: “Which patients do you consider poor AVF candidates?” (See Summary of
Findings, appendix D )
|
|
Nephrologist Pre-Meeting (73% responded) |
Nephrologist Post-Meeting (58% responded) |
Surgeon.
Pre-Meeting
(93% responded) |
Surgeon. Post-Meeting (87% responded) |
|
||||||
|
Attendees were asked to check all patient
types they believed to be poor candidates
- pre and post meeting |
|
||||||||||
|
Geriatric/Elderly |
7 |
|
0 |
|
1 |
|
0 |
|
|||
|
Cardiovascular. Disease |
2 |
|
2 |
|
1 |
|
1 |
|
|||
|
Prior Failed Access |
5 |
|
0 |
|
1 |
|
1 |
|
|||
|
Drug Abusers |
9 |
|
2 |
|
8 |
|
2 |
|
|||
|
Diabetic |
9 |
|
2 |
|
1 |
|
0 |
|
|||
|
Obese |
5 |
|
1 |
|
5 |
|
2 |
|
|||
|
Long term Steroid Use |
4 |
|
0 |
|
3 |
|
1 |
|
|||
|
Other Diseases |
10 |
|
5 |
|
5 |
|
4 |
|
|||
|
Patient Pref/Aesthetic |
0 |
|
1 |
|
0 |
|
0 |
|
|||
|
Short Life Expectancy |
6 |
|
4 |
|
5 |
|
4 |
|
|||