Development of the Medicare
Patient Safety Monitoring
System
Susan L. Abend, MD, FACP
David R. Hunt, MD, FACS
Gaston Mbateng, Ph.D.
Nancy Safer, RN, MSN
Janet P. Tate, MPH
Nancy R. Verzier, RN, MSN, CPHQ
Background and
Development Team
• DHHS Patient Safety Task Force
• CMS
– David R. Hunt, MD, FACS
• Qualidigm
– Connecticut Quality Improvement Organization
• Active Collaborators
– Federal Agency Work Group
• AHRQ, CDC, FDA, VA
– Technical Expert Panel
– Computer Science Corporation (Central Data Abstraction Center)
Goal
To determine the incidence of
specific, clearly defined,
hospital-acquired adverse
events within the Medicare
population
Purpose
• Baseline data for CMS national quality
improvement initiatives
– Surgical Care Improvement Project
• Annual data to the National Healthcare
Quality Report
• A method for repeated assessment of
events (tracking and trending) for safety
improvement activities within healthcare
organizations
Design
• Retrospective Cohort Study
• 25,000- 40,000 randomly-selected
discharges/year from the Hospital
Payment Monitoring Program
– Data obtained from medical charts and
Medicare Part A claims database
• HPMP cases
– randomly selected cases from 50 states, D.C.,
Puerto Rico, U.S. Virgin Islands
– sent to CDAC’s to check coding accuracy
MPSMS Definition of
Adverse Event
“An unintended patient harm, injury, or loss
more likely associated with an interaction
with the health care delivery system than
from an attendant disease process.”
• Patient centered
– Focuses on patient experience
– Does not presume to assign severity
• Detects an untoward outcome
• Requires defined healthcare exposure
• Not dependent on cause
– Process malfunction (error, negligence)
– Imperfect technology
Measure Development
Process
• Event and exposure defined
• Boolean algorithm developed to
detect exposure-related event
• Alpha test
• Beta test
• Clinical review
• Production
Measure Selection
Criteria
• Findable/Feasible
• Adverse event(s) very likely to be
associated with exposure
• Common (burden on Medicare
population)
• Responsible for serious morbidity
and mortality
• Preventable
Adverse Event Rates
’02 and ‘03 Medicare Inpatients
Rate (%) + 95%CI
Post-Operative Pneumonia
2.5 + 0.3
Post-Operative Venous Thromboemboli
0.6 + 0.1
Post-Operative Urinary Tract Infection
3.7 + 0.3
Knee Replacement Adverse Events
7.2 + 1.3
Hip Replacement Adverse Events
11.4 + 1.8
Adverse Event Rates
’02 and ‘03 Medicare Inpatients
Rate (%) + 95% CI
Ventilator Associated Pneumonia
Hospital-Acquired Bloodstream Infection
11.9 + 1.8
0.30 + 0.05
CVC-Associated Bloodstream Infection
1.4 + 0.3
CVC-Associated Insertion Site Infection
2.6 + 0.4
CVC-Associated Mechanical Adverse
Events
2.3 + 0.3
Average Length of Stay
’02, ‘03 Medicare Patients With Invasive
Surgical Procedures
30
with event
25
no event
DAYS
20
15
10
5
0
Post Op
Pneumonia
Error bars indicate 95% CI
Post Op VTE
Post Op UTI
Inpatient Mortality Rate
’02, ‘03 Medicare Patients With Invasive
Surgical Procedures
25
with event
no event
20
%
15
10
5
0
Post Op
Pneumonia
Error bars indicate 95% CI
Post Op VTE
Post Op UTI
Phase II Measures
One Year of Data Pending
• Postoperative Cardiac Events
• Adverse Drug Events
– Anticoagulant-related hemorrhagic events
– Insulin/oral hypoglycemic agent-related hypoglycemic
events
– Antibiotic-associated C. Difficile infection
• Pressure Ulcers
Phase III Measures
In Development
• In- Hospital Falls
• Angiography-Related Adverse Events
– Contrast nephropathy
– Adverse events related to femoral artery
puncture
• Urinary Tract Infections Associated with
Bladder Catheterization
Reliability
Agreement Rate of Reported MPSMS Variables
100
90
dichotomous
80
categorical
70
%t
60
50
40
30
20
10
0
>=95%
>=90% and < 95%
>=80% and < 90%
Kappa Statistic for Reported
MPSMS Variables
> 0.8
0.61 - 0.8
<= 0.6
15%
55%
30%
Limitations
• Validation
– No true gold standard yet defined for
determining sensitivity and specificity
• Data source issues
– Retrospective
– Depends on consistent documentation of
exposures and events in medical record
• Events are relatively infrequent
• Limited capability for risk adjustment
Policy Implications
• Complements reporting efforts
– Able to detect denominator
– No reporting bias
• Complements indicators derived from
administrative data alone
– Potential for improved sensitivity
– Richer database of variables
Policy Implications
cont’d
• Standard definitions
– trackable over time
• Allows hospitals to use common, standard
benchmarks
• Potential for post-discharge surveillance of
events
• Usable in paper or electronic health records
– Potential for concurrent or prospective use in error
trapping or event avoidance.
Summary
• The MPSMS is a reliable tool for measuring adverse
events in hospitalized patients
• Can measure the outcome of errors, suboptimal
systems and/or technologies
• Uses standardized, patient-oriented definitions of
adverse events
• Development is transparent and consensus-driven
– strong collaboration between public and private stakeholders
Conclusion
The MPSMS is a valuable tool for
hospitals and health care organizations
to use in making transformational
changes to create a safe healthcare
environment.
This material was prepared by Qualidigm, the Medicare Quality Improvement Organization
for Connecticut, under contract with the Centers for Medicare & Medicaid Services (CMS), an
agency of the U.S. Department of Health and Human Services. The contents presented do
not necessarily reflect CMS policy. Pub. # QUALCT-PSMS-200501