Develop a plan for observations of provider behavior as they use the PSA CDSS and answer the following questions:What would you look for? (2%)How many observations would you do? (2%)What methods would you use (number of observers, use a predefined tool)? (2%)What kind of data will you have from your observations? (2%)How will you analyze and report the data? (2%)
psa_cdss_proposal__2_.docx
Unformatted Attachment Preview
A Primary Care Provider Intervention to Target PSA-Based
Prostate Cancer Screening
SPECIFIC AIMS
Prostate specific antigen (PSA)-based prostate cancer screening is prevalent in the US,
and many men receive screening against current recommendations.1-4 Approximately
half of men age 75 to 79 years old receive PSA screening1 as do nearly a third of those
with limited life expectancies.2 Most men who undergo PSA screening receive annual
tests, though there is a lack of evidence that this screening interval improves outcomes. 5
Once diagnosed, the majority of men with prostate cancer undergo treatment regardless
of age or health status.6 The resultant morbidity, psychological impact, and financial
costs are high.7
Primary care provider (PCP) recommendations are the single largest predictor of
whether a patient undergoes PSA screening.1,8 PCPs, however, frequently choose
screening without discussing it with patients, have been shown to inaccurately estimate
the likelihood of prostate cancer and benefits of screening, and have received little
guidance on how to help patients weigh the risks and benefits of screening.9,10
In the primary care setting, it is critical to work with PCPs to (1) highlight instances where
existing guidelines agree that routine PSA screening should not be offered, (2)
emphasize scenarios where recommendations have come to differing conclusions
regarding the reasonableness of offering PSA screening (e.g. men age 55 to 69 years
old with 10 year life expectancies), and (3) discuss men for whom there are substantial
gaps in existing evidence (e.g. among black men and those with a family history of
prostate cancer). Tools that help clinicians more accurately estimate life expectancy
are necessary for identifying patients that fall into the first and second categories. For
patients in the second and third categories, PCPs need better information, tools, and
communication techniques to promote shared decision-making about the risks and
benefits of PSA screening with their patients at the point of care.
Using a targeted educational campaign along with a computer-based clinical
decision support system (CDSS) integrated into the PCP workflow, we aim to
better inform the approach to PSA screening. The CDSS will be tested in the Johns
Hopkins Community Physicians (JHCP) practice, the largest outpatient provider in
Maryland with 26 outpatient sites, 172 PCPs, and 40,000 men age 40 years and over
who are eligible for screening. The specific hypothesis is that a CDSS incorporated into
the electronic medical record (EMR) combined with provider education will affect
provider behavior, improve the balance of benefits and harms, and reduce the costs of
PSA-based screening for prostate cancer when compared with current practice.
Specific Aim 1. Develop a CDSS and curriculum for PSA-based prostate cancer
screening
A. Develop a PSA screening algorithm that incorporates best evidence and existing
guidelines
B. Integrate the PSA screening algorithm as a CDSS into the EMR-based workflow
surrounding health care provision
C. Design a curriculum for primary care providers to be used in conjunction with the
CDSS.
Specific Aim 2. Implement a CDSS for PSA-based prostate cancer screening with
education in JHCP
A. Implement the intervention (education plus CDSS).
B. Educate primary care providers, institute CDSS, and monitor utilization over a
two intervention period
Specific Aim 3. Evaluate the impact of a CDSS for PSA-based prostate cancer
screening with education
Informing PCPs’ approach to PSA screening is both critical and innovative. The proposal
offers an important focus on PCP education and decision support; it is strengthened by
its multidisciplinary team of investigators; and it will be tested with a gold-standard
randomized controlled trial in a large, real-world practice setting.
RESEARCH STRATEGY
A. SIGNIFICANCE
In May 2012, the US Preventive Services Task Force (USPSTF) released final
recommendations discouraging the use of prostate specific antigen (PSA)-based
prostate cancer screening for all men based on an assessment that benefits did not
outweigh harms.11 Yet, there is evidence from randomized trials that a subset of men
benefit from screening through a reduction in prostate cancer mortality.10,12-15 An
alternative to discouraging screening for all men is a more targeted approach that
reduces rates of screening among men who are least likely to benefit, and allows
patients a role in weighing the potential benefits and harms of screening based on their
values and beliefs.9,16
Despite differences in recommendations as to whether PSA screening
should be offered to younger men, it is clear that screening rates should ideally be
highest for men in good health in the target age range. The strongest evidence that
benefits may outweigh harms of screening is for the target age group 55-69 years.13
Data from the Rotterdam section of the ERSPC show no benefit (prostate cancer
mortality reduction) for men age 70 years and above for those screened versus not. 15
However, population based studies have demonstrated that PSA screening rates were
highest (56%) for men age 75-79 and were similar (24%) for men age 50-54 and those
age 85 and older.2,17 Among men 70 and over, over 30% of men with limited life
expectancies received PSA screening.2 Current patterns of screening are unlikely to be
fully explained by patient preferences and informed decision-making.8,18-25
For men, diagnostic work-up for an elevated serum PSA concentration may
result in complications and anxiety. Men with an elevated serum PSA concentration
frequently undergo a prostate biopsy.26 Rates of infectious complications following
biopsy have increased in recent years.27-29 Men with an elevated serum PSA
concentration frequently report high distress and anxiety, and this distress may continue
after a negative biopsy.30 Distress may extend to family members and other loved
ones.31-33
Once diagnosed with prostate cancer, the majority of men undergo
treatment regardless of age or health status,6 and treatment may result in
significant side effects, reductions in quality of life, and financial costs. In a recent
study of men with low risk prostate cancer, over half of men with a 70% probability of
death from another cause at 10 years underwent treatment.34 Similarly, among VA
patients age 85 and older, 77% of men diagnosed with prostate cancer received active
treatment.35 Active treatment of low risk prostate cancer has increased with the
introduction of advanced technology even though there has been growing awareness of
active surveillance as a management option.36 Treatment may lead to side effects
including urinary, bowel, and erectile dysfunction in an estimated 20-40% of men37 which
may have negative impacts on health related quality of life for patients and their
families.38-42 Further, the financial costs of prostate cancer diagnosis and treatment are
also high. Prostate cancer costs exceed $12 billion annually and account for nearly 10%
of the total cost of cancer care to Medicare,43,44 and the rapid adoption of new
technologies for prostate cancer treatment has resulted in considerable additional
national spending in the US.45
Among men who undergo PSA screening, the optimal frequency of
screening that balances the risk of over-diagnosis with frequent screening, versus
the potential of missing an aggressive prostate cancer with infrequent screening
is not known. Annual PSA testing is a strategy often recommended if screening is
accepted by patients46, but annual screening has been predicted to result in increased
false positive tests and over diagnosis without substantially improving the benefits of
screening. As compared to biennial screening that was common in the control arm of the
PLCO, annual screening in the intervention arm did not result in a reduction in prostate
cancer mortality. A 2-4 year screening interval used in the European Randomized Study
of Screening for Prostate Cancer (ERSPC) was associated with a reduction in prostate
cancer-specific mortality.10 Computer simulations evaluating different rescreening
intervals suggest that most of the benefits of screening will be preserved with biennial
screening –an approach that would reduce both over diagnosis and over treatment.
Racial and sociodemographic disparities exist for prostate cancer
screening. African-American men are more likely than Caucasian men to have an
advanced prostate cancer at diagnosis, and more likely to die of the disease.47 The
extent to which this disparity is due to differences in screening or management after
diagnosis is not known, but racial/ethnic disparities for PSA screening have been
documented.48-50 Given the similarity in stage distribution among African-American and
Caucasian men in an equal-access health care system (Department of Defense),51 it is
likely that screening practices explain some of the health outcomes differences. It is
plausible that an approach to screening African Americans that differs from Caucasians
may be needed to reduce the racial disparity in outcomes.
Primary care providers play a critical role in current patterns of PSA
screening. First, their screening recommendations are the primary driver of PSA
over-screening among the elderly and those in poor health.1,8 Prior research has
further found that PCPs have been shown to inaccurately assess patient life expectancy
and incorrectly estimate prostate cancer risk and benefits.22,52-54 Existing evidence
suggests that providers may have a limited understanding of tools to estimate life
expectancy and be reluctant to engage in conversations around prognosis. In our prior
survey of Johns Hopkins Community Physician providers (see Preliminary Studies
below), providers additionally expressed concerns over medicolegal risk and lack of time
to discuss with patients.55 Though providers view patient expectations as a barrier to
discontinuing screening, research suggests that these expectations around PSA
screening may be modifiable with a clear provider input.
Second, among patients who may benefit from PSA screening, shared
decision-making is rarely employed. Tools to improve informed decision-making
have focused on improving patient knowledge regarding PSA screening.56,57 However,
informed decision-making often implies that providers have the requisite knowledge and
skills to assist a patient in weighing the risks and benefits of a decision. Many providers
order testing without an informed discussion,8,18-24 and even when patients are informed
about screening, they tend to receive an unbalanced message that emphasizes the
benefits rather than the harms of screening. These data suggest that provider education
through a decision support tool that provides the practitioner with clear guidance on
existing points of consensus and controversy in current guidelines along with the risks
and benefits of screening may support informed decision-making, and reduce the harms
of screening.16
Clinical decision support systems (CDSS) coupled with provider education
and outreach may be an important means to optimize PSA screening. CDSS’s
interfaced with the electronic medical record (EMR) have the potential to improve
medical practice by aligning evidence and practice.58-61 Computerized reminders at the
point of test ordering have been shown to reduce the rate of unnecessary testing,62 and
3 in 4 trials evaluating CDSS’s to reduce unnecessary health care utilization reported
improvements.58 While the effect size of CDSS tools is often small, with a recent review
showing that the median absolute improvement in outcomes was 5.6%,63 reviews
highlight multiple evidence-based approaches that make CDSS more likely to improve
outcomes.64 In particular, CDSS’s that are supported by provider education and outreach
may improve their impact.63,65 With respect to prostate cancer, a tool for combining
evidence from current guidelines and giving providers estimates of risks and benefits for
individual patients may be useful for complex decision-making that involves tradeoffs,66
and providers have highlighted the need for clinical reminder systems for prostate
cancer screening.62,67 Further, education around estimating prognosis in clinical practice
and providing techniques to encourage shared-decision making are essential.68
In summary, results from randomized trials and population-based studies upon
which the USPSTF based their recommendations, and the recommendations of other
professional organizations, may be used to inform PCPs about the potential risks and
benefits of PSA-based prostate cancer screening for specific patients, leading to a more
evidence-based and preference-sensitive approach to PSA-based screening. Current
patterns of screening suggest that such an approach is urgently needed.
B. INNOVATION
There are multiple innovative aspects to the proposed project. Most importantly,
the proposed project seeks to change the current paradigm of a “one size fits all”
approach to prostate cancer screening and reduce the harms of over-screening.
To accomplish this, we will develop, implement, and evaluate a clinical decision
algorithm and support system (CDSS) combined with provider education. Though
CDSSs are frequently used to improve rates of cancer screening, they have not, to our
knowledge, been used to help reduce the rate of inappropriate cancer screening
recommendations or to intercalate into PCP discussions and decision-making.
Specifically, the proposed CDSS will, for individual patients, (a) reinforce decisions
where existing professional societies, expert panels, and best evidence agree that PSA
should not be offered, (b) provide tools to more easily incorporate life expectancy into
clinical decision-making, (c) highlight areas of disagreement and uncertainty among
existing recommendations, (d) provide data that supports shared decision-making, and
(e) among patients who elect screening, improve screening intervals. The approach
builds on existing literature on CDSS that shows that these are most likely to be effective
when they are automatically delivered at the point of clinical decision-making, and when
they provide recommendations that are concrete.64 It is situated within a broader
educational implementation framework designed to support shared decision-making.
From a decision-science perspective, the project will advance the field in a
number of ways. First, it will address the issue of decision support not to take action, the
opposite of most DSSs whose goals are to promote action. Second, it will investigate the
acceptability and use of explicit life expectancy in supporting clinician decision making.
Third, it will deal with recommendations based on evidence and those based on less
evidence in the same DSS, providing a model for other decision domains with a similar
epistemological mix.
Importantly, we partner with a large network of community providers to
implement the intervention in a real-world setting. The setting—Johns Hopkins
Community Physicians (JHCP)—includes a diverse patient population in urban and
suburban practice locations who see over 40,000 men who are eligible for PSA
screening each year. Because we are using the most widely used commercial EHR,69
the CDSS, if successful, may be incorporated into EMRs and replicable in other clinical
settings.
C. APPROACH
Our approach, as outlined in Figure 1, seeks to overcome contributors to PSA overscreening and improve shared decision-making though provider education coupled with
clinical decision support system. In aim 1, we will develop an evidence-based algorithm
that will form the backbone of our education and decision-support tool (Figure 2). We
Figure 1. Overview of proposed research approach
will embed the algorithm within a CDSS to be used by PCPs on an outpatient basis at
the point of care of PSA testing and urological referral. The tool will be integrated into the
clinical workflow and will be implemented as a decision support module of an EMR for
use during the clinical encounter. The CDSS will be constructed through an iterative
process in which we combine patient and population-level evidence with elicited expert
consensus through a Delphi panel, obtain provider feedback, and elicit patient
perspectives. Similarly, we will develop provider education that supports the use of the
CDSS and addresses contributors to PSA screening in conjunction with patients and
providers. Education will further assist with estimating life expectancy in clinical practice
and promote techniques for shared decision-making.
In aim 2, we will implement the CDSS with JHCP primary care providers. We will
use a robust education component combined with frequent monitoring to ensure
adequate uptake.
Setting. JHCP comprises 26 outpatient sites in 11 counties in Maryland. With 179
primary care providers—including 156 physicians, 17 nurse practitioners, and 6
physician assistants—and 225,000 patients, it is the largest outpatient provider network
in Maryland. JHCP practices, which are fully paperless and use e-prescribing, have
successfully meet the CMS Meaningful Use of Electronic Medical Records incentives
criteria. The organization has an internal Information Services division with 17
employees including database analysts, programmers, and hardware specialists. JHCP
currently uses the Epic EMR system, which is widely employed in health systems across
the United States. JHCP has past experience developing CDSS for asthma care, and
preventive care in both adult and pediatric populations, and data on JHCP patients has
been regularly used to support other research studies and performance improvement
projects.70
Research team. We have assembled a collaborative group of investigators with
expertise in 1) the development of CDSS for incorporation into the EMR and adoption by
users (Drs.Lehmann, Marsteller), 2) stakeholder surveys, education, training, and
implementation (Drs. Marsteller71-74, Boyd, Pollack55,75,76) 3) population-based cancer
outcomes (Drs. Pollack,77,78 Carter27,79-82), 4) prostate cancer screening (Drs. Carter,27,8392
), 5) informatics (Drs. Lehmann93-96 and Zeger), and 6) statistical analyses (Dr. Zeger97103
). Members of the team have experience working on the preliminary studies
(described below), completed papers, and on-going grants as described in the
biosketches.
PRELIMINARY STUDIES SUPPORTING THE APPROACH. We have performed
multiple preliminary studies that both lay a foundation for the current work and
demonstrate our team’s ability to effectively accomplish the tasks outlined in this project.
PSA Screening Rates in Johns Hopkins Community Physicians. We have
examined current patterns of PSA screening among men who receive their primary care
at JHCP using EMR data. From April 2012 to April 2013 (following the USPSTF
recommendations against routine screening), there were 39,444 men over age 40 years
without a history of prostate cancer who had at least one primary care provider visit. Of
the 14,880 men ages 55 to 69, 37.4% received PSA screening. Out of 5,669 men age
70 and over, 23.4% received screening. Over 1% (406 men) were diagnosed with
prostate cancer.
Survey of Johns Hopkins Community Physicians. We administered a written
survey of practitioners in JHCP in November 2011.55,76 The survey was conducted at an
annual provider retreat and a $10 conditional incentive was employed. 141 physicians
and nurse practitioners who deliver primary care for adult male patients attended the
retreat and were eligible for participation, and 88.7% responded to the survey. Providers
rated their practice patterns as follows: 17.2% typically order a PSA test without
discussing the harms and benefits with patients, 33.6% discuss the harms and benefits
but generally recommend the test, 43.4% discuss the harms and benefit and then let the
patient decide, 4.1% generally recommend against the test, and 1.6% do not discuss
and do …
Purchase answer to see full
attachment
Recent Comments