Académique Documents
Professionnel Documents
Culture Documents
Ross Hammarstedt
Deborah Bulger
performance improvement
a “left brain meets right brain” approach
Clinical analytics enables hospitals to combine clinical and financial data in developing better
strategies for performance improvement.
In 1979, Betty Edwards introduced the Not surprisingly, most of us develop a dom- lack of community services, patient com-
world to a new drawing technique in her inant mode of thinking based on what is plications, and higher-than-expected
book Drawing on the Right Side of the Brain. more comfortable for us, which results in severity and conclude that the organization
The book, which shot to The New York Times lopsided problem-solving skills. needs more staff. The answer may lie some-
bestseller list within two weeks and stayed where in the middle, or better yet, may be a
there for more than a year, became popular Before flipping to the table of contents to matter of optimizing capacity and through-
for its easy-to-learn approach to drawing, make sure you’re reading the right publica- put by properly aligning existing resources.
which Edwards contends anyone can learn. tion, consider how these two modes of pro- Doing so requires developing a shared per-
It’s a matter of bringing out the abilities of cessing information might apply to the spective based on the ability to tie clinical
the right side—the “creative” side—of the financial and clinical sides of a hospital. outcomes to financial outcomes, also
brain, she says. Financial analysts focus on cost, budgets, known as clinical analytics.
credit rating, revenue, and patient days to
In addition to teaching millions of people determine the bottom line. Clinicians focus Developing a Shared Perspective
with no apparent talent how to draw, on safety, compliance, and outcomes to Clinical analytics involves the application of
Edwards helped to popularize a new brain determine the quality of patient care. business intelligence systems to health
theory. To develop her drawing technique, Before the advent of pay for performance, care. Global corporations have been com-
she drew on the words of scientist and neu- there was very little overlap between these peting on such systems since the 1970s,
rosurgeon Richard Bergland (The Fabric of two perspectives. relying on them to mine vast stores of raw
Mind, New York: Viking Penguin, Inc., data and transform complex relationships
1985, p. 1): Take, for example, a hospital on the path to into easy-to-use metrics.
quality improvement whose average length
You have two brains: a left and a right.
of stay for heart failure and shock is run- The sophisticated mining and reporting
Modern brain scientists now know that
ning 1.25 days above benchmark. Given that power needed to compete in this way
your left brain is your verbal and rational
same information, a CFO and a chief nurs- requires a vendor-neutral data warehouse
brain; it thinks serially and reduces its
ing officer will typically draw very different that extracts, aggregates, and normalizes data
thoughts to numbers, letters and words
conclusions as to why. The CFO will look at from multiple transactional repositories.
... Your right brain is your nonverbal and
patient days, billing delays, claim denials, Rather than replicating entire systems, the
intuitive brain; it thinks in patterns, or
and deferred admissions and conclude that data warehouse extrapolates only data rele-
pictures, composed of ‘whole things,’
the organization needs more beds. The vant to that organization. Business logic is
and does not comprehend reductions,
CNO will see discharge planning failure, then applied that enables the organization to
either numbers, letters, or words.
hfm DECEMBER 2006 I
intimately understand its customers’ buying results, incision time and so on—are cre- language. These scorecards establish an
habits, for example, and personalize its ated during the care process and never used early warning system for clinical and finan-
products and services to maximize profit. again. When combined with other data cial variances, enabling stakeholders to
Data can be sliced, diced, and presented to points and imbued with clinical knowledge, head off errors and drive process improve-
stakeholders in the form of web-based they can provide great insight. ment.
scorecards that enable them to drill down
into key metrics and quickly draw conclu- The exhibits below and on page 103 illus- A good scorecard focuses attention on only
sions. trate the incremental value of combined the most meaningful metrics for that stake-
clinical and financial data. Suppose you holder. Physicians and other clinicians
Clinical analytics requires a similar want to determine the impact of periopera- must be able to see a clear connection to
approach: extracting clinical, financial, and tive care on the cost of a total hip replace- their daily practice. A nurse manager, for
operational data from multiple repositories ment. Relying on financial data alone, you example, would be interested in metrics
and loading it into a warehouse optimized could determine whether the patient was related to barcode medication administra-
for reporting and root cause analytics. Many charged for an anti-infective and a glucose tion compliance on her unit that provide
hospitals overtax the reporting features of test on the day of the surgery—information ongoing feedback regarding whether care is
their clinical repositories, which are useful only from a cost perspective. Clinical improving as a result of barcode scanning
designed to manage real-time clinical work- data alone will allow you to determine, (e.g., whether her nurses are responding
flow related to individual patients during through time stamps, whether the prophy- appropriately to alerts). Metrics are mean-
episodes of care, not for reporting beyond ad lactic anti-infective and glucose test were ingless as ends in themselves; their value is
hoc queries to help clinical supervisors administered according to evidence-based in helping stakeholders know where to
organize care. In contrast, data warehouses guidelines and how long the surgery lasted. drive process changes that will ultimately
are designed to aggregate data retrospectively This information is somewhat more useful. be reflected in higher scores.
and help the enterprise achieve desired out- Only by combining financial and clinical
comes for populations of patients across a data and applying embedded clinical Creating a metric-driven culture also
continuum of care. Embedded clinical knowledge, however, can we get at what we requires data integrity. Besides being
knowledge, or healthcare-specific business really want to know. Using clinical analyt- expensive and time-consuming to collect,
logic, transforms the normalized data into
actionable information.
Clinical analytics provides the long-elusive
Five Ways to Compete on Quality
Clinical analytics provides the long-elusive
transparency needed to directly
transparency needed to directly link care
and cost, thus introducing a tremendous
link care and cost, thus introducing
competitive advantage. As a result, organi- a tremendous competitive advantage.
zations can compete on quality in five
important ways. ics, we can determine the degree to which manual data provide only a snapshot in
procedure duration and noncompliance time for a sample of the population. Manual
Clinical analytics enables clinical performance with guidelines impact readmissions, total data are generally gathered to meet specific
to be measured using clinical data. For more cost of care, percentage of postoperative reporting requirements or answer specific
than 20 years, we’ve measured performance infections, and other important measures, questions. Should a new requirement arise
using cost accounting, largely because such as complications by payer and patient or the data gathered beg a new question, it’s
charge codes with Uniform Bill-92 billing satisfaction. back to the charts. In contrast, clinical ana-
codes have been readily available and fairly lytics has a rapid refresh rate because data
inexpensive. However, the inadequacy of Clinical analytics supports a metric-driven cul- are regularly feeding the warehouse. This
administrative data to provide insight into ture. enables ongoing trend analysis for an entire
the quality and appropriateness of care, A key characteristic of high-performing population. It can also eliminate objections
including errors of omission or commis- organizations is the ability of every about sample size when a given clinician’s
sion, has long been acknowledged. Clinical employee to articulate the meaning of qual- performance is questioned. As new ques-
analytics taps data produced as a byproduct ity in their organization and the impact they tions arise, the data can always be re-
of patient care rather than as a consequence have on quality. Via scorecards, the quality mined.
of patient billing. Thousands of data imperative can be communicated deep
points—on medication administration, lab within the organization using a common Finally, scorecards should enable users to
⫽
Hospital = 98 Physician = 76 Age = 56 > % post op infections when procedure
Payer = Blue Cross duration > 90 minutes
⫹ ⫹ > Complications by payer
Medication Administration Data ⫹ Surgery IS > Patient satisfaction tied to periop care
Rocephin 1 GM 10:04 Left hip replacement
Incision time 10:45
⫹ Close time 12:55
Lab Results
Glucose, fasting 175 mg/dl 06:00
Deborah Bulger
is vice president, product market-
ing, performance management
solutions, McKesson Provider
Technologies, Hadley, Mass.
(deborah.bulger@mckesson.com).