1

Measuring Patient Safety and Disparities using the Medicare Patient Safety Monitoring

System (MPSMS)

9th Annual Maryland Patient Safety Center Conference, April 5, 2013

Noel Eldridge, MSAgency for Healthcare Research and Quality

Center for Quality Improvement and Patient Safetynoel.eldridge@ahrq.hhs.gov

301-427-1127

1

MPSMS Background

Key Aspects of the Medicare Patient Safety Monitoring System (MPSMS)g y ( )– Based on chart review by abstractors at CDAC 21 specific measures

– Patient charts come from RHQDAPU and IQR Abstraction with MPSMS tool comes after other

reviews

– Sample was: all-diagnosis (18+) Medicare for 2002-2007

All payer (18+) 4-principal diagnosis groups (AMI, HF, SCIP, Pneumonia) for 2009-2011,

Potentially “global” (18+) starting in 2012 2

2

MPSMS’s 21 Measures (unofficial breakdown)

Adverse Drug Events

– IV Heparin

L M l l W i ht

AEs during or after surgery or invasive procedures

Mechanical AEs associated– Low Molecular Weight Heparin

– Warfarin

– Insulin and Oral Hypoglycemics

– Digoxin

Hospital-Acquired Infections– CAUTI

– Mechanical AEs associated with CVCs

– Post-op VTE

– Contrast Nephropathy

– AEs associated with femoral artery puncture for angiography

– AEs after Hip surgeryCAUTI

– CVCBSI

– VAP

– C. difficile

– MRSA

– VRE

– Post-op pneumonia

– AEs after Knee surgery

– Post-op cardiac events

Others Pressure Ulcers

Falls

3

Role of MPSMS data in HHS’s Partnership for Patients

MPSMS supplies over 90% of the measured Hospital-Acquired Conditions (HACs) that we are using to track q ( ) gnational changes from 2010 thru 2013

ADEs, CAUTIs, CLABSI, Falls, Pressure Ulcers, VAP, & VTE (7 of 9 focus areas) will be measured at the national level with MPSMS

SSIs are based on CDC’s NHSN data

Obstetric AEs are based on AHRQ PSIs

Additional MPSMS and PSIs are used to represent “all-other” HACs in the estimate

Baseline is 145 HACs per 1,000 discharges in 2010

– based on MPSMS, PSI, NHSN and HCUP data4

3

PFP-Measured National HACs Baseline Pie Chart (2010)

Percent of Total Measured HACs –PFP 2010 Baseline (4.745M)

Adverse Drug Events (57% Hypoglycemic Events & 42% Anticoagulant Drug Events)Pressure Ulcers

34.2%

5 5%1.7%

0.8%

0.4%0.3%

18.8%

Catheter-Associated Urinary Tract Infections

Falls

Surgical Site Infections

Obstetric Adverse Events

27.8%

8.4%

5.5%2.0% Ventilator-Associated Pneumonia

Central Line-Associated Bloodstream Infections

Venous Thromboembolism

All Other HACs -- based on 14 other specific measures (from C diff Infection to Contrast Nephropathy) 5

National Measurement –Hospital-Acquired Conditions (HACs)

We have good measures for the 9 HACs but they are not perfect; forHACs, but they are not perfect; for example…– VTE measure is post-op only

– ADE measures miss many ADEs – e.g., those due to narcotics or allergies, etc.

– CAUTIs or CLABSIs caused by inpatient fprocesses but manifesting post-discharge cannot

be counted

– SSI measure covers only 17 selected procedures

– OB measures count injuries to mother only

6

4

Disparities in 2002-2007 data(background)

Publication by Metersky, et al (including Moy) in Medical Care (May 2011) covered 2004 2007Medical Care (May 2011) covered 2004-2007 data

Design and Subjects: Abstraction of 102,623 Medicare charts of non-Hispanic white and black patients to assess frequency of patient safety events in 4 domains: – adverse drug events due to anticoagulants and

hypoglycemic agents,

– selected nosocomial infections,

– selected procedure-related adverse events,

– general (pressure ulcers and falls) 7

Disparities in 2002-2007 data(summary findings)

Blacks had a higher risk than whites of suffering one of the nosocomial infections (1 34; 95%one of the nosocomial infections (1.34; 95% confidence interval, 1.17-1.55; P < 0.001) and one of the adverse drug events (1.29; 95% confidence interval, 1.19-1.40; P < 0.001).

Patients in hospitals with the highest percentages of black patients were at increased risk of experiencing one of the nosocomialinfections (1.9% vs. 1.5%; P < 0.001) and adverse drug events (8.7% vs. 7.8%; P < 0.01).

8

5

A quick look at potential disparities in 2010 -2011 data

Preliminary analyses of all measures– Years and types of events not aggregated

Measured grouped as follows: ADEs, HAIs, Surgery & Procedures, General

Two sets of slides for each grouping of measures:measures:– Exposure Rates

– Adverse Event Rates among those exposed

9

ADE “Exposure” Rate(Preliminary Data)

This is a measure of the percentage of patients that are subject to the AE, i.e., mostly a measure of those that “get” the drug.

15%

20%

25%

30%

35%

40%

45%

50% Exposure rate (%) WHITE PATIENTS 2010 (12-month, n=27528)

Exposure rate (%) WHITE PATIENTS 2011 n=27699

Exposure rate (%) BLACK PATIENTS 2010 (12-month, n=4088)

Exposure rate (%) BLACK PATIENTS 2011 n=4086

(Preliminary Data)

0%

5%

10%

ADE Associated with Digoxin

ADE Associated with Hypoglycemic

Agents

ADE Associated with IV Heparin

ADE Associated with LMWH and Factor

Xa Inhibitor

ADE Associated with Warfarin

Exposure rate (%) "OTHER" PATIENTS 2010 (12-month, n=2062)

Exposure rate (%) "OTHER" PATIENTS 2011 n=2182

10

6

Adverse Drug Event Rate(Preliminary Data)

This is a measure of the percentage of patients that experience an ADE (as defined by MPSMS) – only among those “exposed”.

4%

6%

8%

10%

12%

14%

16%Observed adverse event rate (%) WHITE PATIENTS 2010 (12-month, n=27528)

Observed adverse event rate (%) WHITE PATIENTS 2011 n=27699

Observed adverse event rate (%) BLACK PATIENTS 2010 (12-month, n=4088)

Observed adverse event rate (%) BLACK PATIENTS 2011 n=4086

* (Preliminary Data)

0%

2%

ADE Associated with Digoxin

ADE Associated with Hypoglycemic

Agents

ADE Associated with IV Heparin

ADE Associated with LMWH and

Factor Xa Inhibitor

ADE Associated with Warfarin

Observed adverse event rate (%) "OTHER" PATIENTS 2010 (12-month, n=2062)

Observed adverse event rate (%) "OTHER" PATIENTS 2011 n=2182

* Statistically significant difference (95% CI) for both years

11

Hospital-Acquired Infection “Exposure” Rate (Preliminary Data)

This is a measure of the percentage of patients that are subject to the HAI, as defined by the MPSMS measure.

(Preliminary Data)

40%

50%

60%

70%

80%

90%

100%

Exposure rate (%) WHITE PATIENTS 2010 (12-month, n=27528)

Exposure rate (%) WHITE PATIENTS 2011 n=27699

Exposure rate (%) BLACK PATIENTS 2010 (12-month, n=4088)

Exposure rate (%) BLACK PATIENTS 2011 n=4086

Exposure rate (%) "OTHER" PATIENTS 2010 (12-month, n=2062)

Exposure rate (%) "OTHER" PATIENTS 2011 n=2182

0%

10%

20%

30%

40%

ADE Hospital Acquired Antibiotic Associated C diff

Blood Stream Infections

Associated with Central Venous

Catheters

Catheter Associated

Urinary Tract Infections

Hospital Acquired MRSA

Hospital Acquired Vancomycin

Resistant Enterococcus

Postoperative Pneumonia

Ventilator Associated Pneumonia

2011 n=2182

12

7

Hospital-Acquired Infection Adverse Event Rate (Preliminary Data)

This is a measure of the percentage of patients that experience an HAI (as defined by MPSMS!) – only among those “exposed”.

(P li i4 00%

6.00%

8.00%

10.00%

12.00%

14.00% Observed adverse event rate (%) WHITE PATIENTS 2010 (12-month, n=27528)

Observed adverse event rate (%) WHITE PATIENTS 2011 n=27699

Observed adverse event rate (%) BLACK PATIENTS 2010 (12-month, n=4088)

Observed adverse event rate (%) BLACK PATIENTS 2011 n=4086

Observed adverse event rate (%) "OTHER" PATIENTS 2010 (12-month, n=2062)

Observed adverse event rate (%) "OTHER" PATIENTS 2011 n=2182

(Preliminary Data)

0.00%

2.00%

4.00%

ADE Hospital Acquired Antibiotic

Associated C diff

Blood Stream Infections

Associated with Central Venous

Catheters

Catheter Associated

Urinary Tract Infections

Hospital Acquired MRSA

Hospital Acquired Vancomycin

Resistant Enterococcus

Postoperative Pneumonia

Ventilator Associated Pneumonia

13

Surgical and Invasive Procedure “Exposure” Rate (Preliminary Data)

This is a measure of the percentage of patients that are subject to the AE, i.e., mostly a measure of those that “get” the procedure.

(Preliminary Data)

15%

20%

25%

30%

35%

Exposure rate (%) WHITE PATIENTS 2010 (12-month, n=27528)

Exposure rate (%) WHITE PATIENTS 2011 n=27699

Exposure rate (%) BLACK PATIENTS 2010 (12-month, n=4088)

Exposure rate (%) BLACK PATIENTS 2011 n=4086

Exposure rate (%) "OTHER" PATIENTS 2010 (12-month, n=2062)

Exposure rate (%) "OTHER" PATIENTS 2011 n=2182

0%

5%

10%

Femoral Artery Puncture for

Catheter Angiographic Procedures

Hip Joint Replacements

Knee Joint Replacements

Contrast Nephropathy

Associated with Catheter

Angiography

Mechanical Complications

Associated with Central Venous

Catheters

Postoperative Cardiac Events for Non-Cardiac

Surgeries

Postoperative Cardiac Events

for Cardiac Surgeries

Postoperative Venous

Thromboembolic Events

14

8

Surgical and Invasive Procedure Adverse Event Rate (Preliminary Data)

This is a measure of the percentage of patients that experience an AE (as defined by MPSMS) – only among those “exposed”.

20%

(Preliminary Data)

4%

6%

8%

10%

12%

14%

16%

18%

20%Observed adverse event rate (%) WHITE PATIENTS 2010 (12-month, n=27528)Observed adverse event rate (%) WHITE PATIENTS 2011 n=27699

Observed adverse event rate (%) BLACK PATIENTS 2010 (12-month, n=4088)Observed adverse event rate (%) BLACK PATIENTS 2011 n=4086

Observed adverse event rate (%) "OTHER" PATIENTS 2010 (12-month, n=2062)Observed adverse event rate (%) "OTHER" PATIENTS 2011 n=2182

15

0%

2%

4%

Femoral Artery Puncture for

Catheter Angiographic Procedures

Hip Joint Replacements

Knee Joint Replacements

Contrast Nephropathy

Associated with Catheter

Angiography

Mechanical Complications

Associated with Central Venous

Catheters

Postoperative Cardiac Events for Non-Cardiac

Surgeries

Postoperative Cardiac Events

for Cardiac Surgeries

Postoperative Venous

Thromboembolic Events

Pressure Ulcer and Falls Adverse Event Rate (Preliminary Data)

In MPSMS 100% of patients are considered “exposed “ for Pressure Ulcers and Falls. Pressure Ulcers are of all stages (I to IV), and Falls i l d ll f ll ( t j t f ll ith i j )include all falls (not just falls with injury).

(Preliminary Data)

3.0%

4.0%

5.0%

6.0%

Observed adverse event rate (%) WHITE PATIENTS 2010 (12-month, n=27528)Observed adverse event rate (%) WHITE PATIENTS 2011 n=27699Observed adverse event rate (%) BLACK PATIENTS 2010 (12-month, n=4088)Observed adverse event rate (%) BLACK PATIENTS 2011 n=4086

Observed adverse event rate (%) "OTHER" PATIENTS 2010 (12-month, n=2062)Observed adverse event rate (%) "OTHER" PATIENTS 2011

0.0%

1.0%

2.0%

Hospital Acquired Pressure Ulcers In-hospital Patient Falls

Observed adverse event rate (%) OTHER PATIENTS 2011 n=2182

16

9

Discussion Items

MPSMS disparities studies are constrained by the small sample size forconstrained by the small sample size for racial/ethnic subpopulations– And many measures only pertain to even

smaller subgroups, e.g., CVC patients, knee surgery patients, warfarin patients, etc.

– Combining multiple years of data for severalCombining multiple years of data for several related measures (e.g., all ADEs) reduces the ability to see trends, but may provide more insights re disparities (as was done for 2004-2007) from 2010 to date

17

Trying to breakdown “Other”

Hospital-Acquired Antibiotic-Associated Clostridium difficile2011 2010

Race/Ethnicity:Exposed

(#)AE (#)

Rate (%)

95% CIExposed

(#)AE (#)

Rate (%)

95% CI

Hispanic 1362 2 0.15 0.02-0.53 1392 11 0.79 0.4-1.41Asian 449 6 1.34 0.49-2.89 395 1 0.25 0.01-1.4American Indian/Alaska Native 193 0 0 0.00-1.89 112 1 0.89 0.02-4.87Native Hawaiian/Pacific Islander 22 0 0

0.00-15.44 25 0 0

0.00-13.72

Other 607 1 0.16 0.00-0.91 584 1 0.17 0.00-0.952633 2508

18