2. Founded in 1852 Catholic, non-profit Fairly highly
integrated system across 5 westernstates 27 hospitals from critical
access to largequaternary facilities Plus over 100 other kinds of
facilities ~53,000 employees Annual revenue: ~$8.5 billion NOI
typically 4.5-5% ~1500 employed physicians (out of 13,000medical
staff members)
5. People and Culture The ability to instill a culture of
collaboration,creativity, and accountability. (i.e. a
learningorganization that embraces a just culture) Business
intelligence The ability to collect, analyze, and connect
accuratequality and financial data to support
organizationaldecision making. (More on this later)
6. Performance improvement The ability to use data to reduce
variability inclinical processes and improve the delivery ,
cost-effectiveness, and outcomes of clinical care. (More on this
later ) Contract and risk management The ability to develop and
manage effective carenetworks and predict and manage different
formsof patient-related risk. (i.e. integrated ACOs with good
data)
10. Data Re-Use, AKA business intelligence,Unified
Intelligence, Comprehensive dataWarehouse Relatively new concept
for health care Uses still being defined and explored
Quantification of costs are pretty clear Quantification of
benefits: still emerging
11. Reduce the cost of collecting/analyzing data Speed the
decision making process and fasterspread of innovation, based on
near real-timeaccess to information Preclude the need for many,
future smalldatabase acquisitions Anticipate having data to answer
questions thatwe didnt know wed want to ask Identify which data
(among many) really need tobe standardized Reduced waste and injury
Data backupPage12
12. Selected after a look at the options: review by outside
consultants (First, Gartner) site visits to other users (SJHS)
Initial Implementation: 2 of 4 regions 7+ use cases Evaluation of
technical deployment, user friendliness,future use, cost of
ownership Enterprise agreement to support Providencedeployment Pa
ge 13
13. Initial Goal: identify the top 10 uses for
initialimplementation Actually, we stopped after 47 potential uses
Create the supporting infrastructure For managing the tool For
spreading and implementing knowledge acrossthe system Connect as
many data sources as possible. Pa ge 14
14. System implemented in all 4 regions107different data inputs
Governance/communications structure created Support staff
hired--~26 FTEs (mostly internalstaff) Continued training of key
users Initial focus on 8 key uses: Catheter-Associated UTIs
Modified Early warning system (MEWS) Sepsis risk Central line blood
stream infection Readmission trackingPage15
15. Use cases, contd Core measureCHF D/C Patient transfer
activity Glycemic monitoring. Key strategic concept: use system to
identifypatients requiring standardized interventions(but allow
staff to also do ad hoc inquiries)Page16
16. In one (smaller) Providence region, it costs $7million per
year to collect and report coremeasure data Due to brute force data
collection: clinicians go on thewards to find core measure
candidates and hand tallyresults. Amalga solution: Replace brute
force method by using electronic data tofind those patients and
alert ward staff. Expansion of core measures will only increasethe
cost problem if a data warehouse does notexist.
17. A pharmacy alert system Multiple electronic inputs (lab,
pharmacy, ADT) Locally developed rules scan the inputs and
alertpharmacist to intervene with at-risk patients. Impact (at 20
hospitals in 1 system): $4 million pharmacy savings/ month 70
serious events averted in 4 months Amalga could have done this, had
it beenpurchased earlier (and will eventually replace thatsystem)
Caveat: system worked best where pharmacistsworked on the
wards.
18. Improving sepsis outcomes Early detection and treatment of
sepsis: up to 50%mortality improvement; 30% improvement in LOSand
30% improvement in cost of care. So, why stop at treating patients
who already havesepsis? Next step: Use Amalga to ID patients at
greatestrisk for sepsis for intensified monitoring andprevention of
sepsis.
19. Catheter-associated UTIs (CA-UTIs) 50% of HAIs In
Providence, HAIs cost ~$45 million/yr Equivalent to 247 nurses, who
could be put tobetter use Prior to Amalga it was impossible to even
know whohad urinary catheters Now, catheter patients can be
identified andevidence- based standards applied Expected outcome:
50% reduction in CA-UTIs
20. Post MEWS75.465.455.4Pre MEWS45.435.4UCL25.4CL15.4 5.4
LCL-4.61/23/2012
21. 14Total Pages for Code Team from "PEAT" areas12UCL10Post
MEWS &PEAT rounds 8 6CL 4 2 0 1/23/2012
22. $2,700 $453,60014Potential annualsavings using MEWS Average
Admissionsreimbursement 35that couldshortfall for an be avoidedICU
admission (HFMA July 2006) (McQuillan 1998)Monthlyescalations
ofcare to ICUMore importantlyMEWS at PAMC saved lives1/23/2012
23. Basic concept: use data mining to detect bestpractices
within ones own system Internal best practices more likely to be
adopted e.g. Most cost effective approach to stroke,pneumonia, hip
replacement Also, need appropriate communicationssystem and
infrastructure to support spreadand adoption Ultimately, more rapid
adoption of innovationmeans faster savings, improvements.
24. New clinical registriesfor relating inputs tooutcomes
Orthopedics Thoracic surgery Real-time ICU dashboards More active
data mining by qualitydepartment. Incorporation of cost data into
the system.
25. This is a new toolnot intuitive to many folks Communicate,
communicate, communicate! Educate, educate, educate! Involve all
stakeholders in the process. This is a pluri-potential tool Know
which strategic goals are key This is an expensive tool Know
(roughly) how its going to pay for itselftellstories to illustrate,
know the costs that can bereduced, even though the actual results
are not inyet; have an idea why real-time data are important. Pa ge
27
26. This is not an expensive tool compared to thevarious
alternatives Even with a state of the art EMR, this kind of
toolmakes sense. This is not a magical tool Understand that it
needs to be supported by a skilledstaff and effective
infrastructure We have probably underestimated its potentialuses
and value It is primarily limited by the amount of electronicdata
available So, we expect that well want to generate even moresources
of electronic dataPage28
27. Access to real-time data reveals multipleopportunities to
improve clinical outcomesAND financial returns In general, the
actual benefits have turnedout to be greater than the estimated
benefits Data alone are not sufficient; also requiredare: Skilled
data mining and presentation Supporting infrastructure to act on
the data