10/22/2018

1

Dashboards, Scorecards, & Performance Metrics

Christopher Caspers, MD, FACEP

Chief, Observation Medicine

NYU Langone Health

Associate Professor

NYU School Medicine

President-Elect

Observation Medicine Section

American College Emergency Physicians

Disclosures

• I have no conflicts of interest or disclosures.

2

10/22/2018

2

3

• Define key observation metrics and learn their importance

• Use outcomes data to identify the best delivery model for observation

care

• Learn how new technology can be used to establish surveillance

systems in observation care

• Build dashboards to improve metrics in observation medicine

• Understand how to synchronize protocolized observation care into the

electronic health record (EHR)

Objectives

Key Metrics

• Volume

• Length of Stay

• Conversion Rate

• Recidivism

• Patient satisfaction

10/22/2018

3

Volume

• The number of patients cared for in the OU, influenced by:

• Inclusion Criteria

• Patients requiring the active management of their condition following the initial ED visit to

determine the need for inpatient admission or discharge

• Exclusion Criteria

• No clear working diagnosis

• No clear management plan

• Acute exacerbation of psychiatric condition

• Acutely altered mental status

• Hemodynamic instability

• Sepsis

• Requirement for nursing evaluation more frequently than every 4 hours

• Agitated, combative or acutely intoxicated patient (may be placed in Observation Services after

clinical sobriety achieved in ED)

Volume

• Monitor by protocol

– Overutilization

– Underutilization

• Evaluate need for new protocol

– ‘General protocol’

10/22/2018

4

Volume → How many beds?

• Size matters

• Clinical breadth

• Simple vs Complex observation

• Resource justification

7

Smaller Units → Less Protocols Relatively Simple Observation

8

Protocols at go-live were a best-guess guess of what our patients would be like

10/22/2018

5

Predictable Daily Ebb and Flow1159pm

Full

overnight

Less full in

early

afternoon

Rapidly

filling in

evening

•OU rapidly fills in

the evening and

overnight

•Majority of

dispositions

occur in the late

morning and

afternoon

1200am

Average Hourly Arrivals/Dispositions

Arrivals decrease -- ‘Down time’ -- Rounding and Dispos -- Arrivals increase.

10/22/2018

6

Unit Structure

Key Metrics

• Volume

• Length of Stay

• Conversion Rate

• Recidivism

• Patient satisfaction

10/22/2018

7

Length of Stay

• Therapeutic protocols have longer LOS

• Troubleshoot prolonged obs stays

– Priority testing

– STAT turnaround time

• Diagnostics

• Labs

• Consults

– Intrinsic factors

• Staffing, workflows

• Patient selection, management

Key Metrics

• Volume

• Length of Stay

• Conversion Rate

• Recidivism

• Patient satisfaction

10/22/2018

8

Conversion Rate

• The percent of patients admitted to inpatient status at the end of OU

care

• Marker of OU efficacy and resource matching

– Goal 15-20%

• Too high: patient selection, workflow issue

• Too low: patient selection, missed opportunities

• Exception:

– Complex observation: trend towards higher rate

Key Metrics

• Volume

• Length of Stay

• Conversion Rate

• Recidivism

• Patient satisfaction

10/22/2018

9

72-Hour Revisit after ED Obs Discharge:

(16% returned to Obs)

72-hr revisit rate after Obs discharge: 4-5%

30-day return to IP rate after Obs discharge: 3.5%

72 Hour Revisit after ED Obs Discharge: Causes of Revisit

10/22/2018

10

Key Metrics

• Volume

• Length of Stay

• Conversion Rate

• Recidivism

• Patient satisfaction

Patient Satisfaction is Higher in a Dedicated Obs Unit

8381

94 95

70

75

80

85

90

95

100

15-bed T1 OU 35-bed T1 OU

Patient Satisfaction Scores (Press Ganey)

Type 4 OU Type 1 OU

Patient satisfaction higher

when observation services are

provided in a Type 1 OU setting.

10/22/2018

11

Patient Satisfaction

Started with all OU discharges. 358 comments

in total.

41 comments were identified as being negative and consequently omitted from qualitative analysis.

116 of the remaining comments contained just a

single word such as “Excellent” or “Good”, just a staff member’s name or

other non-substantive text.

201 of the remaining

positive comments were coded based on theme.

123 comments were in some

way related to hospital staff.

49 comments referenced the staffing overall without

specifying the department. -

“The staff were so caring and kind. I felt very safe.”, “The staff was excellent in every

regard.”

46 comments specifically cited

the nursing staff. – “The nurses were incredible - very

professional and efficient - but comforting

and calm. I had complete confidence in them.”

17 comments cited the

physicians/PAs. – “the PA and the doctor were both

incredible, empathetic and sweet.”

Remaining 11 comments about

staff included those about thosewho drew blood and

performed tests such as CT scans.

78 comments were on their stay overall. - “Can’t think of anything. It’s really first-rate”, “Could not have

been better.”

Dashboards

• Measurement

– Automatic

• Surveillance

– Early recognition

• Key to quality improvement

– PDSA

• Upward and outward management

– Tell your story

22

10/22/2018

12

Dashboards:Clinical Performance

High frequency, high

rate

Protocol performance is continuously monitored to ensure desired

outcomes and to identify opportunities for improvement.

Example: Gastrointestinal Bleed Protocol Inpatient Conversion Rate Too High (57%)

Conversion rate for protocol is higher than expected (57%; goal 15-

20%)

10/22/2018

13

Example: Gastrointestinal Bleed Protocol

• Exclusion criteria: a function of resources, capabilities of unit, desired

outcomes

• GI Bleed Protocol Exclusion Critieria: use data to derive exclusion

criteria, created decision support tools

Gastrointestinal Bleed Protocol – Improved Patient Selection

Conversion rate for protocol corrected and increased volume of

patients on protocol through data-driven inclusion/exclusion criteria.

10/22/2018

14

Venous Thromboembolism Prophylaxis Guidelines

VTE Prophylaxis Ordering

10/22/2018

15

CHF Protocol ‘Problem State’

29

Metric Performance

Volume 87

Length of Stay 35hrs

Conversion Rate 32%

Diuretic dosing 80%

Pathway launch 46%

Follow-up w/HF Cardiologist 32%

Integrating an Evidence-based Clinical Protocol into the EHR

• Prompt providers to select protocol

• Streamline protocolized care

– Time, clinical parameters, etc

• Incorporate safeguards to prevent readmissions

10/22/2018

16

Unit-level Dashboard

• Real-time monitoring of patients on unit

• Rapid identification of patient on HF pathway

10/22/2018

17

CHF Protocol After Pathway Intervention

33

Metric Pre Post

Volume 87 62

Length of Stay 35hrs 36hrs

Conversion Rate 32% 17%

Diuretic dosing 80% 85%

Pathway launch 46% 95%

Follow-up w/HF Cardiologist 32% 90%

Thank you!

Christopher Caspers, MD

Chief, Observation Medicine

NYU Langone Health

christopher.caspers@nyumc.org