Embed Size (px)
Citation preview
ORIGINAL ARTICLE
Quality Improvement Measures Lead to Higher SurveillanceRates for Hepatocellular Carcinoma in Patients with Cirrhosis
Fasika B. Aberra • Mary Essenmacher •
Natalie Fisher • Michael L. Volk
Received: 23 June 2012 / Accepted: 8 October 2012 / Published online: 1 November 2012
� Springer Science+Business Media New York 2012
Abstract
Background Cirrhosis is a major risk factor associated
with the development of hepatocellular carcinoma (HCC).
The American Association for the Study of Liver Diseases
recommends surveillance for HCC in cirrhosis patients
with ultrasound every six months. However, various stud-
ies suggest that surveillance rates in actual practice are
quite low.
Aim The aims of this study were to evaluate the effec-
tiveness of implementing quality improvement (QI) mea-
sures in increasing the rate of HCC surveillance among
patients in a tertiary care facility.
Methods Patients with cirrhosis were prospectively
enrolled into a chronic disease management program,
which integrates nursing-based protocols with automatic
reminders when patients are due for surveillance. Patients
enrolled in this program between March 2010 and April
2011 were compared to a prior cohort in 2008–2009. The
primary endpoint was the receipt of at least one abdominal
imaging study performed for the purposes of surveillance
during the study period.
Results Of the 355 patients enrolled, 331 (93 %) had
imaging performed for HCC surveillance, compared to
119/160 (74 %) patients in the previous cohort
(p \ 0.001). Chart review revealed the most common
reasons for failure to undergo surveillance were patients’
lack of insurance and lack of follow-up on studies ordered
at outside institutions. Six patients were diagnosed with
HCC during the study period, of which three were at early
stage.
Conclusions Implementation of QI measures incorporat-
ing automatic reminders of surveillance status for providers
can significantly increase the rate of HCC surveillance
among cirrhosis patients.
Keywords Hepatocellular carcinoma � Cirrhosis �Surveillance � Quality improvement
Introduction
Hepatocellular carcinoma (HCC) is the fifth most common
cancer worldwide and the third most common cause of
cancer-related mortality in the United States [3], where
there has been a rapid increase in the incidence rate over
the past four decades [1, 4]. The majority of HCC
(*70–90 %) arises in the setting of underlying chronic
liver disease and cirrhosis [1, 2, 5]. The prognosis of
patients with HCC is dependent on the stage of the tumor at
the time of presentation [6]. Resection, liver transplantation
and percutaneous ablation are shown to achieve a high rate
of complete response and improved survival rates in
patients with early stage disease [7].
Surveillance for HCC in high risk patients can facilitate
early detection and thus increase the chance of diagnosis at
a stage when the tumor is amenable to curative treatment
[8]. The American Association for the Study of Liver
Diseases practice guidelines suggest that surveillance
F. B. Aberra � M. Essenmacher � N. Fisher � M. L. Volk
Division of Gastroenterology, Department of Internal Medicine,
University of Michigan, Ann Arbor, MI, USA
e-mail: [email protected]
N. Fisher
e-mail: [email protected]
M. L. Volk
e-mail: [email protected]
F. B. Aberra (&)
141 Neese Dr. # C18, Nashville, TN 37211, USA
e-mail: [email protected]; [email protected]
123
Dig Dis Sci (2013) 58:1157–1160
DOI 10.1007/s10620-012-2461-4
should be performed at 6-month intervals using ultrasound
in all high-risk patients, which primarily includes those
with cirrhosis as well as certain patients without cirrhosis
who have hepatitis B. AASLD also recommends that sur-
veillance be conducted in the setting of a program in which
standardized surveillance tests and recall procedures along
with quality control measures are in place [9]. However,
reports from previously conducted studies in Veterans
Administration hospitals and the Surveillance, Epidemiol-
ogy, and End-Results-Medicare database showed only
22–28 % of patients with cirrhosis received regular sur-
veillance for HCC [10, 11]. Our previous study showed that
even in a tertiary care hepatology practice, only 74 % of
patients had received HCC surveillance at least once per
year [12].
These findings suggest that deficits exist not only in the
realm of provider knowledge, but also in systems of care.
Most practices rely upon providers ordering the surveillance
ultrasound when patients are physically in clinic, which
might be forgotten while addressing acute issues for that
visit. Then, the test has to get scheduled, the patient needs to
show up to the appointment, and the results need to be
transmitted back to the clinic. Therefore, we instituted a
series of quality improvement (QI) measures aiming to
increase rates of HCC surveillance. The purpose of this study
is to describe these efforts and analyze their effectiveness.
Methods
Patient Population
Patients seen at our hepatology clinic were prospectively
identified by an attending physician as having cirrhosis.
The diagnosis of cirrhosis was based on histology or
imaging showing a cirrhotic appearing liver with associ-
ated signs of portal hypertension including splenomegaly,
varices, or thrombocytopenia. Hepatitis B-infected patients
who have not developed cirrhosis and patients with per-
sonal history of HCC were excluded.
Quality Improvement (QI) Program
During the planning phase, we recognized that improving
rates of HCC surveillance would require several steps.
First, eligible patients would need to be routinely identi-
fied. Second, the ordering of surveillance would need to be
detached from the actual clinic visit. Third, a reminder
system would need to be established. To accomplish step 1,
a box was added on the clinic checkout form for the
attending physician to indicate that the patient has cirrho-
sis. Clerical personnel were trained to notify a designated
staff member when this box was checked. This staff
member was then charged with entering the patient into a
software program designed for chronic disease manage-
ment (Avitracks, Avicenna Inc.), and providing the patient
with educational material about HCC surveillance.
To accomplish step 2, a protocol was written and agreed
upon by all physicians to allow ultrasound surveillance
tests to be ordered per protocol by nursing staff without
seeking physician approval. The software program pro-
vides a reminder for liver ultrasound and alpha fetoprotein
(AFP) monitoring every six months. Physicians were still
able to modify the order at their discretion, such as
obtaining MRI or CT imaging if patients were noted to
have a suspicious lesion on ultrasound or elevated AFP.
Finally, to accomplish step 3, the disease management
software was programmed to alert nursing staff when any
patient was more than one month delinquent on his or her
surveillance. These alerts could be modified or turned off
as appropriate. Implementation of these measures was
complete by March 2010.
Statistical Analysis and Comparison Group
The primary endpoint of the study was the proportion of
patients who underwent surveillance, defined as at least one
abdominal imaging test in the prior year performed for sur-
veillance purposes. Imaging was determined to be for sur-
veillance in cases where this was explicitly stated in
associated clinical notes or as the indication on the radiology
requisition [12]. For the purpose of this analysis, we included
patients that were enrolled in the program over a 1-year
period between April 2010 and May 2011. The surveillance
rates for the current cohort (after QI implementation) were
compared to our previously reported rates (before QI
implementation) in 2008–2009. The pre- and post-QI cohorts
were matched in terms of median age (56 and 57, respec-
tively), male gender (60.6, 58 %), and white race (81.7,
85 %). The etiologies of cirrhosis were also similar, with
viral (42 and 42.5 %, respectively), alcohol (14.4, 13 %),
NASH (16.9, 11 %) and Other (26.2, 34 %) for the pre-QI
and post-QI cohorts respectively. Of note, the pre-QI cohort
consisted of patients who agreed to participate in a research
study, and thus may be inherently more likely to adhere to
surveillance. As patients in the present cohort were part of a
systematic QI initiative, waiver of informed consent was
allowed by our Institutional Review Board.
Demographic and clinical characteristics of the present
cohort were gathered, including age, sex, race, etiology of
liver disease, and Child–Turcotte–Pugh score. The etiology
of liver disease was categorized as viral (hepatitis B or C),
alcohol induced, nonalcoholic steatohepatitis, or other
etiologies.
T test was used to compare the mean age between the
surveillance and non-surveillance groups in the cohort.
1158 Dig Dis Sci (2013) 58:1157–1160
123
Chi-square test for homogeneity was used to compare the
proportion of patients in the present cohort who received
surveillance to the proportion previously reported. Chi-
square test was also used to analyze patient characteristics
associated with receiving surveillance. Statistical signifi-
cance was defined as a P value less than 0.05. All data
analysis was performed using SPSS version 19.
Results
Between April 2010 and May 2011, 362 patients with cirrhosis
were enrolled into the surveillance program. Out of 362
patients, seven patients were excluded from the data analysis
because of personal history of HCC. Clinical and demographic
characteristics of the remaining 355 are shown in Table 1. The
majority of the patients were well compensated with 71 %
(n = 253) having Child–Pugh Class A cirrhosis. During the
study period, 331 (93.2 %) of the 355 patients received
imaging surveillance. This represents a 19 % improvement
when compared to the pre-QI cohort (p \ 0.001).
In sensitivity analysis, after excluding 21 patients with
Class C cirrhosis, the rate of surveillance remained the
same at 93.6 %. There was no statistical difference in the
mean age for surveillance versus no surveillance group (58
versus 55; p = 0.17). The most common surveillance test
used was ultrasound (n = 312), with 19 patients having
MRI performed at the physician’s discretion. None of the
patients had CT scan for surveillance for the duration of the
study period. Review of the charts of patients without
surveillance revealed the most common reasons to be lack
of insurance coverage and lack of follow-up for tests
ordered at an outside institution. During the study period,
323/355 (91 %) patients had alpha-fetoprotein levels
measured, including 19 out of the 24 patients who had not
received imaging surveillance.
Out of the 331 patients undergoing surveillance, 6
(1.8 %) were diagnosed with HCC during the study period.
Three of these were unifocal and less than 3 cm, and thus
early stage. None of the 24 patients without surveillance
were diagnosed with HCC during the study period.
Discussion
This study has demonstrated that QI measures can result in
improved rates of surveillance for HCC among patients
with cirrhosis. Over a period of 3 years, we improved
surveillance rates from 74 to 93 %. During this period there
were no changes in physicians or nurses; it was only the
mechanics of care delivery that changed. Furthermore,
these changes were relatively simple and inexpensive to
implement. Therefore, we conclude that these QI efforts
were successful.
Despite this success, several logistical problems remain.
Although no new nursing or clerical staff were hired,
implementation of the program required more staff time
than initially anticipated. We did not achieve 100 % sur-
veillance, and medical record review of delinquent cases
suggests that the tests were often ordered to be done at an
outside institution. It is unclear whether the patient failed to
adhere to testing, or the test was done but results never
transmitted. Tracking down outside results consumes a
significant amount of nursing and clerical time; for this
reason, we have begun scheduling ultrasounds to be done
in conjunction with clinic visits. Conversely, patients
occasionally have duplicate tests ordered by the primary
care physicians and/or local gastroenterologists, and this
unnecessary duplication contributes to high healthcare
costs. These issues highlight the need for improved com-
munication and coordination of care between providers.
Our study has several important limitations. It was not a
randomized trial, so we cannot be certain that other tem-
poral changes contributed to the increase in surveillance
rates. The comparison cohort (pre-QI) was self-selected by
their consent to participate in a research study, while the
Table 1 Characteristics of
patients
HBV indicates hepatitis B virus,
HCV hepatitis C virus, NASHnon-alcoholic steatohepatitisa Expressed as median (range)
Variable All patients
(n = 355)
Imaging surveillance
(n = 331)
No imaging surveillance
(n = 24)
p
Age (years)a 57 (21–89) 57 (21–89) 55 (22–77) 0.39
Male sex, n (%) 204 (58) 192 (58) 12 (50) 0.44
White race n (%) 301 (85) 280 (85) 21 (88) 0.70
Etiology of liver disease n (%) 0.45
Viral (HCV or
HBV)
149 (42) 138 (42) 11 (46)
Alcohol 46 (13) 45 (14) 1 (4)
NASH 40 (11) 36 (11) 4 (17)
Other 120 (34) 112 (34) 8 (33)
Child Pugh
(% class A)
253 (71) 237 (72) 16 (67) 0.61
Dig Dis Sci (2013) 58:1157–1160 1159
123
post-QI cohort included all patients with cirrhosis in our
clinic. However, this methodological issue should, if any-
thing, bias the study towards the null hypothesis (no dif-
ference between the groups). It is also conceivable that
patients in the pre-QI cohort who were compliant would
continue to be compliant with surveillance in the post-QI
cohort. However, the proportion of overlap between the
current and the previous cohort could not be assessed since
we no longer have identifying information on patients from
the previous cohort. Hence, it is difficult to assess whether
QI would improve adherence to surveillance in patients
who had originally not complied with recommendations.
Our experience in an academic tertiary care practice cannot
necessarily be generalized to other healthcare settings,
though we suspect that even larger improvements might be
seen in settings with lower baseline rates of surveillance.
Including high risk groups other than cirrhosis patients in
the study might also increase the surveillance rate. Finally,
future studies should ideally measure whether such
improvements in processes of care lead to improved patient
outcomes.
In conclusion, we found that QI measures lead to higher
surveillance rates for HCC. This finding demonstrates the
importance of focusing on the mechanics of healthcare
delivery, not just individual provider knowledge.
Conflict of interest None.
Financial disclosures None.
References
1. El-Serag HB. Epidemiology of hepatocellular carcinoma. ClinLiver Dis. 2001;5:87–107.
2. El-Serag HB. Hepatocellular carcinoma: an epidemiologic view.
J Clin Gastroenterol. 2002;35:S72–S78.
3. El-Serag HB, Davila JA, Peterson NJ, Mcglynn KA. The
continuing increase in the incidence of hepatocellular carcinoma
in the United States: an update. Ann Int Med. 2003;139:817–823.
4. El-Serag HB. Hepatocellular carcinoma: recent trends in the
United States. Gastroenterology. 2004;127(5 Suppl 1):S27–S34.
5. El-Serag HB. Hepatocellular carcinoma. N Engl J Med. 2011;
365(12):1118–1127.
6. Pons F, Varela M, Llovet JM. Staging systems in hepatocellular
carcinoma. Hpb (Oxford). 2005;7:35–41.
7. Llovet JM, Schwartz M, Mazzaferro V. Resection and liver
transplantation for hepatocellular carcinoma. Semin Liver Dis.
2005;25:181–200.
8. Zhang BH, Yang BH, Tang ZY. Randomized controlled trial of
screening for hepatocellular carcinoma. J Cancer Res Clin Oncol.2004;130:417–422.
9. Bruix J, Sherman M, American Association for the Study of Liver
Diseases. Management of hepatocellular carcinoma: an update.
Hepatology. 2011;53:1020–1022.
10. Davila JA, et al. Use of surveillance for hepatocellular carcinoma
among patients with cirrhosis in the United States. Hepatology.
2010;52:132–141.
11. Davila JA, et al. Utilization of screening for hepatocellular car-
cinoma in the United States. J Clin Gastroenterol. 2007;41:
777–782.
12. Singal AG, et al. Patient involvement in healthcare is associated
with higher rates of surveillance for hepatocellular carcinoma.
J Clin Gastroenterol. 2011;45:727–732.
1160 Dig Dis Sci (2013) 58:1157–1160
123
