Ensuring Remote Telemetry Monitoring Patient Safety · Remote Telemetry Monitoring ... – Brain...

Ensuring Remote Telemetry Monitoring Patient Safety

IIE/Society for Health Systems

February 2006

Terry Dunn

Health Alliance of Greater Cincinnati

Feb 2006

Session Objectives

You will learn...

1. How to create a specific and meaningful definition of safety.

2. How to build an FMEA consistent with JCAHO requirements.

3. How to measure performance against the safety definition.

4. How to select and pilot safety solutions.

5. How to include, promote, and educate stakeholders to ensure buy-in.

Feb 2006

Overview

• Defining Safety

• FMEA’s & JCAHO

• Measuring Safety Performance

• Selecting & Piloting Solutions

• Strategies to Effect Change

• Summary

Feb 2006

Background

• Numerous studies prove: respiratory & cardiac failure is almost always presaged by observable physiologic changes 6-8 hours before the event, which can be detected for most patients by monitoring of heart rate & O2 saturation levels.

• Effective interventions for cardiac and respiratory failures must begin within 3-5 minutes to prevent long-term negative outcomes:– Death

– Brain damage

– Other systemic injury

• House-wide telemetry eases bed assignments

• Centralized (vs. unit-based) model less costly

Feb 2006

“The Right Thing to Do”

• National Patient Safety Goals (2004) – Improve effectiveness of patient care alarms

• Assure alarm systems that monitor patients are regularly tested and adjusted, if needed, to prevent any problems.

• Alarms are turned on with the correct settings and are loud enough to be heard within the patient unit.

• IHI 100,000 Lives Campaign– Rapid Response Teams – monitoring a signal when

RR appropriate!

Feb 2006

The Problem at University

• Vendor’s alpha site for centralized system—no up-front expectation of potential failures & no support to identify benchmarks

• Nurses at the hospital were often not responding to alerts AT ALL, let alone in a timely manner…– Patient deaths had occurred (though fortunately not recently!)

– Near misses were occurring regularly…

• Because of preventable system failures, some clinical events were not being detected

• Repeated examination of the problem had surfaced numerous root causes, but without significant, widespread, or lasting impact

• Project chartered under recommendations of the Board to fulfill JCAHO requirements

Feb 2006

Defining “Safety”

• Who?

• What?

• When?

• Where?

• SO WHAT?

• Why?

• What can we do about it?

Tips for Success:• Use brainstorming & fishbone

diagrams to build consensus

• Good “starters”:• What does “safe” look like?• What has to happen to

ensure safety?

• Watch out for hidden agendas!!

• Keep the focus on safety, NOT on feelings & grudges…

Tips for Success:• Use brainstorming & fishbone

diagrams to build consensus

• Good “starters”:• What does “safe” look like?• What has to happen to

ensure safety?

• Watch out for hidden agendas!!

• Keep the focus on safety, NOT on feelings & grudges…

Feb 2006

Think Systems!

Technology

Hand-offs

Communication

Processes

Shared Understanding& CommonVision

Feb 2006

Project Title:TUH CMU - Improving Safety of Centrally

Monitored Telemetry Patients

Business Case:

Objectives/Goal Statements:

Business Y: Scope: All TUH nursing units supported by CMU.

Champion: R. Fischer Assistant CNO

Process Owner: K. Ghaffari Director Pt Care Services

Finance Rep: Define 20-Feb-05 4-Feb-05 11-Feb-05 Complete

Black Belt: T. Dunn Measure 15-Mar-05 7-Feb-05 1-Apr-05 In process

Analyze 15-Apr-05 30-Apr-05

Improve 1-Jun-05 1-Jun-05

Control 15-Jul-05 15-Jul-05

Project Team: K. Bally, A. Crawford, K. Ghaffari, J. Goetz, F. McNeil

Problem: 85-90% of CMU alerts receive either no

acknowledgement from the nursing unit or response

times exceed current policy specifications.

K. Ghaffari Director Pt Care Services, A. Crawford

Clinical Mgr CMU/6S, S. Hoell Clin Mgr 6NW, W. Lowe

Clin Mgr 9CCP

Steering Team:

Timeline:

Orig Plan

End

Actual

Start

Current

End

2. Reduce acknowledgement defect rate by 70%. 1. Reduce technical alerts by 70%.

Timely & appropriate intervention to cardiac events is strongly

correlated to successful outcomes. Response to early indications can even prevent cardiac arrests. To optimize the

safety of patients monitored by CMU, we must ensure (1) the

right patients are (2) monitored using individualized

parameters (3) by a reliable system to whose alerts we respond in a (4) timely and (5) clinically appropriate manner.

Safety starts with a focused assessment, properly maintaining the equipment, and effectively communicating alerts and

responses.

Status

Volume of codes (& volume of code-related

deaths)

By June 15, 2005:

Charter DMAIC

Defining Safety:To optimize the safety of patients monitored by CMU, we must ensure:

(1) the right patients are monitored (2) using individualized parameters (3) by a reliable system to whose alerts (4) we respond in a timely and

(5) clinically appropriate manner.

Safety starts with a focused assessment, properly maintaining the equipment, and effectively communicating alerts and responses.

Defining Safety:To optimize the safety of patients monitored by CMU, we must ensure:

(1) the right patients are monitored (2) using individualized parameters (3) by a reliable system to whose alerts (4) we respond in a timely and

(5) clinically appropriate manner.

Safety starts with a focused assessment, properly maintaining the equipment, and effectively communicating alerts and responses.

Objectives/Goal Statements:

2. Reduce acknowledgement defect rate by 70%.

1. Reduce technical alerts by 70%.

By June 15, 2005:

Note that this assumes a far more

comprehensive definition of

“system” than the 2004 National

Patient Safety Goals!

Note that this assumes a far more

comprehensive definition of

“system” than the 2004 National

Patient Safety Goals!

Feb 2006

Overview

• Defining Safety





• Summary

Feb 2006

JCAHO Standard PI.3.10

• An ongoing, proactive program for identifying unanticipated adverse events and safety risks to patients is defined and implemented.

– Rationale: Hospitals should proactively seek to identify and reduce risks to the safety of patients. Such initiatives have the obvious advantage of preventing adverse events rather than simply reacting when they occur. This approach also avoids the barriers to understanding created by hindsight bias and the fear of disclosure, embarrassment, blame, and punishment that can happen after an event.

References: 1. JCAHO CAMH Update 3, Aug 20052. Rich, D.S., Complying with the FMEA

Requirements of the New Patient Safety Standards, MS Powerpoint presentation, JCAHO 2001

References: 1. JCAHO CAMH Update 3, Aug 20052. Rich, D.S., Complying with the FMEA

Requirements of the New Patient Safety Standards, MS Powerpoint presentation, JCAHO 2001

Feb 2006


• Elements of Performance:

1. Select a high-risk process to be analyzed• At least one high-risk process is chosen annually—the

choice should be based in part on information published periodically by the Joint Commission about the most frequent sentinel events and risks

2. Describe the chosen process (for example through the use of a flow chart)

3. Identify the ways in which the process could break down or fail to perform its desired function (i.e., what engineers call potential “failure modes”)

4. Identify the possible effects each failure mode could have on patients and the seriousness of those effects

Feb 2006


• Elements of Performance: (continued)

5. Prioritize the potential failure modes

6. Determine why the potential failure modes could occur

7. Redesign the process and/or underlying systems to minimize the risk of the effects on patients

8. Test and implement the redesigned process

9. Monitor the effectiveness of the redesigned process

Comments: 1. Note that it never says “FMEA”, but it describes the process!2. When you speak to JCAHO representatives, they can’t offer

any examples other than FMEA’s!

Comments: 1. Note that it never says “FMEA”, but it describes the process!2. When you speak to JCAHO representatives, they can’t offer

any examples other than FMEA’s!

Feb 2006

High Risk Processes

• Processes that involve risks or may result in sentinel events– Medication use

– Operative and other procedures

– Use of blood and blood components

– Restraint use

– Seclusion, when a part of care

– Care/services provided to high-risk populations

– Resuscitation

Feb 2006

Constructing an FMEA Table

1. Flow-chart the process

2. Identify actual AND potential failure points in the process

3. Document each failure point

• What failure can/does occur (“mode”)

• Each separate effect resulting from the failure (“effects”)

4. Evaluate each mode/effect pair

• S = Severity of effect

• F = Approximate frequency

• D = Ability of system to prevent, control, or detect the failure

5. Calculate Risk Priority Number: RPN = S x F x D

Feb 2006

Process Flows

10/10/2005 1

Chart 0.0 - High Level Flow Chart of TUH CMU Telemetry Monitoring Process

Nu

rsin

g U

nit

CM

UD

oct

or

MonitorTelemetry (See

Chart 1.3)

MaintainTransmitter (See

Chart 1.2)*

Respond toAlerts (SeeChart 1.3)

InitiateMonitoring (See

Chart 1.1)

OrderTelemetry

DiscontinueOrder

ReviewReports?

END 0.0

Shift Reports

Alerts

Call CMUCall CMU &

RemoveTransmitter

Pick-upTransmitter*/

**

CleanTransmitter

Complete order should include ptID, type of monitoring, andparameters/ranges to bemaintained (Parameterssometimes are not specified--usestandard ranges.)

* When CMU staff are on a nursing unit, they are supposed to check monitor & attend toany technical problems detected for all monitored pts on that unit.

Sometimes notuntil ptdischarged

** CMU staff are not supposed to remove a transmitter from a patient--only a nurse onthe floor should do so.

Log patientout

DMAIC

Feb 2006

Process Flows

9/21/2005 2

Chart 1.1 - CMU Initiates Telemetry Monitoring

CM

UN

urs

ing

Un

itD

oct

or

Receive Callfrom Nursing

Unit

Runneravailable?

Log pt oncharge list &admit board

Runner selects& programs/

tests transmitter

Enter pt info & parametersinto monitor "sector"

assigned to transmitter #

Runner takes transmitterto patient & places leads

Inform nursing unit ofexpected delay &/or

request they come to pickup transmitter

Runneravailable?

Pull & program/test transmitter

Set transmitteraside for pick-up

END 1.1

No

Yes

No

YesParallel whenrunneravailable

Run admission strip(once placed) &

interpret

9/21/2005 1

Chart 0.0 - High Level Flow Chart of TUH CMU Telemetry Monitoring Process

Nurs

ing U

nit

CM

UD

oct

or

MonitorTelemetry (See

Chart 1.3)

MaintainTransmitter (See

Chart 1.2)*

Respond toAlerts (SeeChart 1.3)

InitiateMonitoring (See

Chart 1.1)

OrderTelemetry

DiscontinueOrder

ReviewReports?

END 0.0

Shift Reports

Alerts

Call CMUCall CMU &

RemoveTransmitter

Pick-upTransmitter*/

**

CleanTransmitter

Complete order should include ptID, type of monitoring, andparameters/ranges to bemaintained (Parameterssometimes are not specified--usestandard ranges.)

* When CMU staff are on a nursing unit, they are supposed to check monitor & attend toany technical problems detected for all monitored pts on that unit.

Sometimes notuntil ptdischarged

** CMU staff are not supposed to remove a transmitter from a patient--only a nurse onthe floor should do so.

Log patientout

9/21/2005 3

Chart 1.2 - CMU Telemetry Transmitter Maintenance

CM

UN

urs

ing U

nit

Doct

or

Pt onMonitoring

Recordcharges

Set up dailylog Enter

charges

Remove & repositiontransmitter when ptoff unit (see chart

1.2.1)

Replace batteriesq24 (q8 if SpO2

monitoring)

Check &replace leads

q24

Troubleshootplacement, batteries,

leads, etc. PRN

Pt onMonitoring

3 days?Stamp MD notes to

ask re: discontinuatin

N

Y

Run strips q8Round q8 tomount strips

in charts

9/21/2005 4

Chart 1.2.1 - CMU Telemetry Transmitter Maintenance - Patient Leaves/Returns to Unit

CM

UN

urs

ing

Un

itD

oct

or

Pt to leave unit fortest/procedure

Dialysis?Transporter

arrives

Transporterasks nurseto removetransmitter

Nurse callsCMU &

removestransmitter

Pt offunit

Pt off unit

Transporterarrives

Transporternotifiesnurse toreplace

transmitter

Nurserepositionstransmitter

Routine monitoringresumes upon

reception of signal

Place "pt off unit"indicator on

monitor sector

Nurse should checkleads & batteries forreplacement

Y

N

9/21/2005 5

Chart 1.3 - Monitor Telemetry / Respond to Alerts

CM

UN

urs

ing

Un

itD

oct

or

Pt onMonitoring

Systemdetects

abnormalcondition

CMU techinterprets

alarm

Speed-dialredphone

on nsg unit

Lethalrhythm?

Enter textmsg & pageRN of record

Logincident

Timelyresponse?

Receivealarm

Assess ptneed

Acknowledgealarm

Log acknow-ledgement

Resolvealarm/need

Repeat/Escalate alert

Continue

Y

N

N

Y

DMAIC

Feb 2006

Kinds of Monitoring AlertsDistribution of Alerts - Before

Clinical

14%

Battery

28%

Leads

48%

Other

10%

DMAIC

Observations:

• Only 1 out of 7 alerts is a clinical issue—and some of these are actually erroneous due to technical problems.

• Battery alerts (28%) are completely predictable & thus preventable.

• Probably ½ or more of the Leads Off alerts (48% of total) are preventable.

• Nearly all of the Other (10%) category are preventable.

• Not shown: 17% of alerts—1 of every 6!—are repeats of previous alerts not addressed!!

Observations:

• Only 1 out of 7 alerts is a clinical issue—and some of these are actually erroneous due to technical problems.

• Battery alerts (28%) are completely predictable & thus preventable.

• Probably ½ or more of the Leads Off alerts (48% of total) are preventable.

• Nearly all of the Other (10%) category are preventable.

• Not shown: 17% of alerts—1 of every 6!—are repeats of previous alerts not addressed!!

Feb 2006

Cause & Effect: Potential Causes Brainstorming

Unsafe

Condition

Equip Problems

Wrong Patients

Poor Response

Poor Info

Nursing assignment notknown

Wrong RN list ed for pager

Wrong pt list ed for t ransmit t er

Pager not worn

Pager doesn't work

Phone doesn't work

Nurse not available t o respond

Too many alarms = not hingimport ant

Proper response for pt notknown

Don't know how t oacknowledge CMU

Response req'mt s not known

Order missed/lost

Delay in placingt ransmit t er on pat ient

Wrong t ransmit t er placedon pat ient

Transmit t er not available

Not ordered for pat ient

Paramet ers not specific forpt needs

Pt noncompliant /uncooperat ive

Monit oring syst em fails

Pt moves t o non-monit oredarea

Leads

Bat t eries

Pts a

wai t in

g di sch

arg

e

moni to

red

DNR p

t s mon

it or e

d

Pt co

ndi t io

n d

oes

n't warra

nt

use

Tra

nsm

it ter s l o

st

Bat

t eri e

s no

t

c hang

ed

rou

t i nel y

Low

si gnal

ignore

d

(role

/resp

.

clari t y)

Too m

any le

ads

Fell o

f f

Gel d

ri ed ou

t

At ta

ched inc

orre

ct l y t o

t ran

smit te

r

Pla

ced in

corre

ct ly on p

t

What could reduce monitored patient safety?

Unsafe

Condition

Equip Problems

Wrong Patients

Poor Response

Poor Info

Nursing assignment notknown

Wrong RN list ed for pager

Wrong pt list ed for t ransmit t er

Pager not worn

Pager doesn't work

Phone doesn't work

Nurse not available t o respond

Too many alarms = not hingimport ant

Proper response for pt notknown

Don't know how t oacknowledge CMU

Response req'mt s not known

Order missed/lost

Delay in placingt ransmit t er on pat ient

Wrong t ransmit t er placedon pat ient

Transmit t er not available

Not ordered for pat ient

Paramet ers not specific forpt needs

Pt noncompliant /uncooperat ive

Monit oring syst em fails

Pt moves t o non-monit oredarea

Leads

Bat t eries

Pts a

wai t in

g di sch

arg

e

moni to

red

DNR p

t s mon

it or e

d

Pt co

ndi t io

n d

oes

n't warra

nt

use

Tra

nsm

it ter s l o

st

Bat

t eri e

s no

t

c hang

ed

rou

t i nel y

Low

si gnal

ignore

d

(role

/resp

.

clari t y)

Too m

any le

ads

Fell o

f f

Gel d

ri ed ou

t

At ta

ched inc

orre

ct l y t o

t ran

smit te

r

Pla

ced in

corre

ct ly on p

t

What could reduce monitored patient safety?DMAIC

Feb 2006

Failure Mode/Effect Rating Scales Used

1

2

Rate Frequency Ability to Detect/Prevent/Control1 Very rare - 1/year or less Almost always caught - "99% predictable"2 Unusual - 1/month Usually detected/acted upon - "85% +"3 "A few times" per month4 Infrequent - 1/week Sometimes missed5 "A few times" per week 50/50 detected/acted upon6 Multiple times weekly7 Common - daily Easily missed; requires attention to detail8 "A few times" per day9 Several times daily Hard to detect - " < 10% "10 Multiple times daily Almost never recognized until too late - "1%"

Rate each impact/result of failure for its severity. The following scale is suggested, the any number between 1 and 10 is fine. If uncertain about an item, choose a higher number.

Rate each potential failure for (a) the frequency with which it occurs and (b) the system's ability to detect or prevent or control that failure. The following scales are suggested, though any number from 1 to 10 is acceptable. If uncertain about an item, choose a

Potential permanent injury

KEY: Higher number = higher risk = bigger issue. If in doubt, be conservative and pick a higher number.

Potentially life-threatening

Immediate risk of death

SeverityNo consequence to pt or staff; nuisanceNo consequence to pt but added work

Potential of missing more critical indicationsPotential acute injury

Assigning three scales of 1-10 results in each M/E pair Risk Priority Number (RPN) ranging

from 1 (=1x1x1) to 1,000 (=10x10x10)

Assigning three scales of 1-10 results in each M/E pair Risk Priority Number (RPN) ranging

from 1 (=1x1x1) to 1,000 (=10x10x10)

Feb 2006

FMEA: Failure Modes & Effects Analysis

Attribute of Safety Potential Failure Impact/Result Severity Freq'cy

Ability to

Detect or

Prevent Risk

Awareness of patient condition

RN not available when alert issued (in procedure or personal needs)

Potentially miss critical clinical indications 4 7 7 196.0

Right patients on Telemetry

Monitoring parameters not individualized to patient's specific needs/condition

Alarms not indicative of patient need 4 8 6 192.0


No beds available on units trained to care for patients with dysrhythmia

Patient placed on unit designated for "telemetry", though staff may be unprepared for patient's

needs4 8 5 160.0

Trust/ Communication

No/slow response to alerts; Nursing staff unaware of potential problem

Pt not monitored; potentially miss critical clinical indications

5 6 5 150.0


Too many non-lethal alerts; Nurses feel CMU irritating

No slow/response = multiple calls & increasing frustration; Potentially ignore or miss critical

clinical indications3 8 6 144.0


Patient off unit & system shows "failure"Pt not monitored; potentially miss critical clinical

indications4 8 4 128.0


Units trained to care for patients with dysrhythmia may be not have enough staff

to take a new patient.

Patient placed on unit designated for "telemetry" because they have more staff available, though

staff may be unprepared for patient's needs4 6 5 120.0


Staff/patients not trained in how to replace probes or leads removed temporarily

Pt not monitored; potentially miss critical clinical indications

3.5 7 4.5 110.3


No/slow response to alerts; CMU feels nurses don't care

Multiple calls & increasing frustration 3 7 5 105.0


No response, leading to escalation Multiple calls & increasing frustration 3 7 5 105.0


System crash Potentially miss critical clinical indications 10 1 10 100.0

Current

Identified 42 possible risks, spanning severity from “annoying” to “immediate risk of death”

Identified 42 possible risks, spanning severity from “annoying” to “immediate risk of death”

DMAIC

Feb 2006

Another look at the RPN scoring…

Pareto Chart of TUH CMU Failure Risks - Baseline FMEA

196

192

160

150

144

144

128

120

105

105

100

100

100

96

96

96

96

90

84

84

80

80

78

78

75

72

60

60

56

50

42

40

36

35

35

32

30

16

12

6

- - - -- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -

50

40

30

24

110

5%

11

%

15

%

19

%

23

%

27

%

31

%

34

%

46

%

49

%

51

%

54

%

57

%

59

%

62

%

64

%

67

%

69

%

71

%

74

%

76

%

78

%

80

%

82

%

84

%

85

%

87

%

88

%

89

%

90

%

91

%

92

%

93

%

94

%

95

%

96

%

96

%

96

%

97

%

99

%

99

%

10

0%

43

%

40

%

37

%

0

50

100

150

200

250

10

30

50

70

90

110

130

150

170

190

210

230

250

270

290

310

330

350

370

Reference # of Failure Mode & Effect Pair

Ris

k =

Severi

ty x

Fre

qu

en

cy x

Dete

cti

on

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

Cu

mu

lati

ve %

of

To

tal R

isk

Addressed

Deferred

Cumltv%

Project will address factors

accounting for 96% of

estimated risk

Though some pairs had low RPN’s, they would be addressed because solutions targeting specific types of failures would reduce or eliminate them. One type of failure was determined

to be outside the scope of this project, accounting for the “lost” 4%.

Though some pairs had low RPN’s, they would be addressed because solutions targeting specific types of failures would reduce or eliminate them. One type of failure was determined

to be outside the scope of this project, accounting for the “lost” 4%.

Feb 2006

Overview

• Defining Safety





• Summary

Feb 2006

Metrics

• Critical to Quality (CTQ) Project Requirements

– 100% detection of unsafe events

– High confidence in validity of alerts

– Timely response to alerts

– No rework

• Operational Metrics Defined:

– Business Metrics• Volume of codes

• Volume of deaths related to codes

– Process Metrics• Technical alerts per monitored patient day

• Acknowledgement defects (i.e., no or slow responses) as % of alerts

– Financial Metrics• No significant cost or revenue changes anticipated

– Consequential Metrics• None identified

DMAIC

Feb 2006

Original Baseline

• Nearly 90% of telemetry alerts are not acknowledged in a timely manner, as outlined by hospital policy– Over 1/2 of alerts receive no acknowledgement *

– Over 30% receive acknowledgements outside of specified times

– Acknowledgement of lethal rhythms only slightly lower defect rate than other: 84% defective vs. 91%

• 10 telemetry transmitters lost (@ $2,500each)

Can we be sure a patient is safe if we’re not sure anyone checks & resolves a detected problem?

Can we be sure a patient is safe if we’re not sure anyone checks & resolves a detected problem?

* This is not saying “patient need not addressed”. Rather, CMU had no feedback from the nurse.

DMAIC

Feb 2006

Overview

• Defining Safety





• Summary

Feb 2006

Conclusions: Key Drivers of Defects

• “Noise” overload: Too many non-critical pages & possible misuse of red phones

• Knowledge deficits: – Transmitters are tied to specific units, programmed for

specific patients

– Acknowledging alerts much easier than most realize!

• Missing/broken processes:– Transmitter maintenance

– Return of transmitters to CMU at discontinuation

– Transport of monitored patients

– Back-up coverage when nurse temporarily not available

– Back-up coverage for system failure

– Incomplete data

– Communication system failure

• Incomplete/unclear orders for telemetry

DMAIC

Feb 2006

Solutions

Solutions target 3 separate groups:

• CMU

• Physicians

• Nursing Unit Staff

DMAIC

Feb 2006

Reducing measurement error

• Analysis of log data revealed incomplete recording of alerts & times

• Appeared defects might have been overstated…

• …but retesting suggested the severity of the problem was actually underestimated!!

• New checklist-based logs implemented to simplify recording & summarization/analysis

DMAIC

Feb 2006

Other CMU Department Changes

• Established on each unit a single location to collect transmitters removed from patients when telemetry discontinued

• Began writing patient name on transmitter pouches, and instituted “double check” procedures to verify new patient set-up

• Developed new orientation & annual competency educational materials for nurses

• Developed departmental procedure for routinely testing red phones

• Discontinuing daily pager tests– Inconsistent use & response

– Duplicates diagnostic analysis available from paging system

DMAIC

Feb 2006

Ordering practices

• New order form– Specification of type of

monitoring needed

– Specification of parameters

• New order process– Form faxed to CMU

– Scheduled for conversion to CPOE

Insert picture of order form here…

ORDCHK

UNIVERSITY HOSPITAL, INC. PHYSICIAN’S CHECKLIST/

ORDER SHEET All applicable orders have been checked. ORDERS NOT CHECKED ARE NOT TO BE FOLLOWED.

Orders are modified according to the medical condition of the patient. All orders are to be dated, timed and signed by a physician. Additional orders may be entered at the end of the order set. If the orders are transcribed in sessions, the transcriber must date, time, and initial in the section marked order noted. If the entire set of orders is transcribed at one time, make a single slash across the page and enter the date, time, and your initials. PAGE 1 OF 1

ALLERGIES: ¨ None Known ¨ Yes, Drug/Reaction: ______________________________________________

ORDER NUMBER ü TELEMETRY ORDERS FOR CENTRALIZED MONITORING

UNIT ORDER NOTED

(DATE/TIME) (INITIAL)

ü Use for patients admitted to 9 CCP, 8 CCP, 5 ST, 7 NW (non-cardiac), 6 NW (non-cardiac), 4 NE

ü If patient needs cardiac dysrhythmia monitoring, please use Telemetry Admission Orders.

ü Attending Physician: Resident: Pager:

ü Diagnosis:

ü Reason for CMU Monitoring:

ü Patient’s Baseline Vitals: HR:______________ 02 Sats:_____________________

0xygen saturation monitoring. Call for saturation less than_________.

Heart rate monitoring. Call for heart rate less than __________ or greater than________.

ü Discontinue telemetry monitoring after 72 hours.

ü IF PATIENT REQUIRES MONITORING BEYOND 72 HOURS, YOU MUST REORDER!!

ü Please fax these orders to the CMU, 584-6681.

CMU Standard Monitoring Parameters (to be used if inclusive of patient’s baseline)

HR Less Than 50.

HR Greater Than 135.

Oxygen Saturation Less Than 90.

White--Chart Yellow--Pharmacy Pink--Floor Copy Physician Signature Date Time_________________________ Developed by Andy Crawford, RN, BSN Date 04/2005 Review Date 04/2006

DMAIC

Feb 2006

Routine Maintenance Critical!!

Date

In

div

idu

al

Va

lue

2/18/20052/14/20052/10/20052/6/20052/2/20051/29/20051/3/200512/30/200412/26/200412/22/200412/18/2004

10

5

0

-5

_X=0.85

UC L=2.68

LC L=-0.98

1-Before 2-A fter

Date

Mo

vin

g R

an

ge

2/18/20052/14/20052/10/20052/6/20052/2/20051/29/20051/3/200512/30/200412/26/200412/22/200412/18/2004

8

6

4

2

0

__MR=0.688

UC L=2.247

LC L=0

1-Before 2-A fter

11

I-MR Chart of LowBatteryN by Period

DMAIC

One unit piloted regular daily battery changes for telemetry

transmitters.Immediately the volume of low

battery alerts from CMU dropped from ~5 to <1 per day!

One unit piloted regular daily battery changes for telemetry

transmitters.Immediately the volume of low

battery alerts from CMU dropped from ~5 to <1 per day!

Feb 2006

Acknowledgement & Maintenance

4/13/2005 1

“How can I get CMU to quit bugging me?!”

“How can I get CMU “How can I get CMU to quit bugging me?!”to quit bugging me?!”

To minimize CMU pages & alerts:To minimize CMU pages & alerts:

1. Fresh batteries in the transmitteru Change q24 for patients on cardiac-only monitoring

u Change q8 for patients on SpO2 monitoring

2. Good lead contact, with proper lead placementu Electrode gel dries out. Check during AM care &

replace PRN, at least q48. Initial & date at least one electrode when replacing the set.

u ALWAYS replace electrodes if pulled off or fallen off.

u Refer to the lead placement diagram on transmitter.

u Remember to put transmitters back on patients returning from tests/other areas.

3. Acknowledge an alert as quickly as you can!u Lethal rhythm/red phone alerts in 1 minute

u Pages in 3 minutes

u To acknowledge a page, just push the button so CMU knows you’re with the patient!

Press & hold the black button for count of 3 to send

acknowledgement signal—NO NEED TO TELEPHONE!!

Press & hold the black Press & hold the black button for count of 3 to send button for count of 3 to send

acknowledgement signalacknowledgement signal——NO NEED TO TELEPHONE!!NO NEED TO TELEPHONE!!

4/13/2005 1


















DMAIC

• “Just press the button” to acknowledge an alert

• Institute “preventive maintenance” routines– Anticipate electrode gel

drying out; replace leads Q48

– Anticipate battery depletion:• Q24 change for cardiac-only

monitoring

• Q8 when monitoring SpO2

• Increase awareness/ ”WIIFM” among nurses– Posters

– Staff communication

– Revamped training• Orientation

• Intranet-based annual competencies

Feb 2006

Transport & Back-up

• Drafted new policy outlining procedure for transporting monitored patients– Transporters reminded that they must check-in with

patients’ nurses before taking a patient

– Reflects community standards Codifies intended (but currently not standard) practice at University

– Will be incorporated into education for nurses, HUC’s, and transporters

• Drafted policy outlining back-up procedures for system failure– Will be incorporated into education for nurses, HUC’s,

and other Patient Care Services personnel

DMAIC

Feb 2006

CALL TO ACTION – July 2005

What unit managers/leaders need to do NOW

• Incorporate battery & lead changes into routine shift vitals

• Remind staff:– All alerts & responses are logged, including who was alerted

– Get a “buddy” to hold your pager when you won’t be available (off unit, break, in procedure, etc.)

– “Just push the button” to acknowledge an alert

– Put discontinued transmitters & logs in the return boxes

• Look for news of next steps:– Policy & procedure updates

– New ordering process

– New educational materials

– Results of remeasurement

What we asked of

managers in July…

What we asked of

managers in July…

DMAIC

Feb 2006

Baseline CapabilityAfter Implementing New Data Collection Forms

* Target: “…70% Reduction…” of original for responses, of new baseline for per-monitored-patient-day statistics

** Z-score: >50% defect, Z <0. “Near perfect” z=6.

Not Applicable

2.270.68Not Available

Technical Alerts per Monitored Patient per Day

Not Applicable

2.63Not Available

Not Available

Alerts per Monitored Patient per Day

-0.7677.6%27%90%

% Alerts Not Acknowledged in Timely Manner

2.01

44.3%

New Baseline

<00.60

Not Available

Defective Alerts per Monitored Patient per Day

0.1415%50%% Alerts Not Acknowledged

Z-score

**

New Target*

OriginalMetric

Some “organic” improvement was expected, as educational posters &

communications were in place before “new baseline” was taken…

Some “organic” improvement was expected, as educational posters &

communications were in place before “new baseline” was taken…

DMAIC

Feb 2006

Results of Communication Audit

• Overall, 90% of questions asked were answered correctly!!

• Of missed answers, 80% from 3 questions:– #18 – Transmitters can be matched to the patient by

comparing the name on the pouch to the patient’s ID

– #19/20 – Dealt with transportation. If going to Dialysis, nurse should call CMU to get a transmitter that will work in Dialysis. Otherwise, the nurse should notify CMU the patient will be off-unit, and leave the transmitter in the patient’s room.

Conclusion: Communication was received—staff know what to do!

Conclusion: Communication was received—staff know what to do!

DMAIC

Feb 2006

CMU: Knowledge Audit of Units Served

Findings by Unit & Position

2620 21

63

14

2619 22

80

9

2619

201

24

36

23 20

12

58

12

25

14 17

80

9

2114

186

21

34

88%

100%

57%

92%

86%

96%

74%77%

100% 100%

81%

74%

93%

88%

94%

0

50

100

150

200

250

A : A

ll

B : A

ll

C : A

ll

D : A

ll

E : A

ll

F : A

ll

G : A

ll

H : A

ll

I : A

ll :

All

: H

UC

All

: P

CA

All

: LP

N

All

: R

N

All

: M

anager

All

: O

ther

Unit : Position

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

Attempted

Correct

% Correct


Manager reported no

communication had occurred

Manager reported no

communication had occurred

DMAIC

Feb 2006

CMU: Knowledge Audit of Units Served

Analysis of Responses by Question

29

28

13

16

17

16 16 16 16

12

16

13

14

13

16

15

14

11

13

11

29

23

12

16

17

13

15

16

13

12

14

12 12

13

15 15

14

6

9 9

100%

82%

92%

100% 100%

81%

94%

100%

81%

100%

88%

92%

86%

100%

94%

100% 100%

55%

69%

82%

0

5

10

15

20

25

30

35

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

Question

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

Attempted

Correct

% Correct

Results of Communication Audit DMAIC

Feb 2006


• Wrong Answers

– Staff were immediately coached/corrected when they provided an incorrect answer

• Comments/Concerns from staff

– Clarity re: responsibilities/assignments

• “Anybody can change the batteries and leads…”

• “Whoever is available changes the batteries and checks the leads.”

– Limited staffing

• “Sometimes CMU says we have to pick up the transmitter, but we only have 2 staff ourselves!”

• “If we have to escort a monitored patient, we’ll leave too many patients on the unit for the rest of the staff…”

Recommendations: 1. Regular transmitter maintenance MUST be defined as part of the

daily routine on all units.2. Transport can be called to get a transmitter picked up.3. The Nursing Supervisor can advise on options—another unit may be

able to spare some help to escort or fill-in.

Recommendations: 1. Regular transmitter maintenance MUST be defined as part of the

daily routine on all units.2. Transport can be called to get a transmitter picked up.3. The Nursing Supervisor can advise on options—another unit may be

able to spare some help to escort or fill-in.

DMAIC

Feb 2006

Results of Remeasure

JVery good!!

Improved 26%

(p=0.05) 1.682.27

Technical Alerts per Monitored Patient per Day

JVery good!!

Improved 28%!!

(p=0.036)1.892.63

Alerts per Monitored Patient per Day

Improved 32%!!

(p=0.015)

Improved 9%

(p=0.01)

No change (p=0.10)

Change

JGood..we can still do

better!

68.5%77.6%% Alerts Not Acknowledged in Timely Manner

JVery good!!

1.362.01

Defective Alerts per Monitored Patient per Day

LWe can do better!

Interpretation

48.7%44.3%% Alerts Not Acknowledged

RemeasureNew Baseline

Metric

Still a long way from goal of 0.70…

Still a long way from goal of 0.70…

DMAIC

Feb 2006


CMU Alerts by Type - Before & After

14% 11%

28%

23%

20%

21%

27%

25%

11%

20%

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

0-Before 1-After

Other

Leads

SpO2

Battery

Clinical

Other than the reduced number of alerts per

patient, what changes occurred in types of

alerts?

The Chi-square test for independence is used to compare distributions. A p-value < 0.05 indicates

the distributions are different (i.e., independent of one another). For the before & after samples, p = 0.99—the distributions are statistically the same!

Other than the reduced number of alerts per

patient, what changes occurred in types of

alerts?

The Chi-square test for independence is used to compare distributions. A p-value < 0.05 indicates

the distributions are different (i.e., independent of one another). For the before & after samples, p = 0.99—the distributions are statistically the same!

DMAIC

Feb 2006


Nearly every unit showed improvement!!Nearly every unit showed improvement!!

Ack

no

wle

dg

em

en

t D

efe

cts

pe

r P

ati

en

t

Unit

Group

CSD

9CC

P

8CC

P

7NW6S

6NW

5NW5E4N4E

1-A

fter

0-B

efore

1-A

fter

0-B

efore

1-A

fter

0-B

efore

1-A

fter

0-B

efore

1-A

fter

0-B

efore

1-A

fter

0-B

efore

1-A

fter

0-B

efore

1-A

fter

0-B

efore

1-A

fter

0-B

efore

1-A

fter

0-B

efore

6

5

4

3

2

1

0

CMU Alert Acknowledgement Defects per Patient per Day by Unit, Before & After

Improvement can be seen when “After” has either:

1. Lower midline/crosshairs.2. Smaller box.

DMAIC

Feb 2006


• Observations from data:

– 17% of alerts were repeats/escalations• Repeats were NOT occurring in 5-minute intervals—15-60

minutes passed between attempts

– Documented incidents:• Nurses took their pagers with them off-unit & did not respond

for hours

• Missed red phone alert (phone tested by communications and no operational problems found)

• Abusive & avoidance behaviors toward CMU when non-acknowledgement followed-up by telephone calls to unit

– Results varied by individuals on same units: some always responded on time, others in same area never responded.

This variability points to individual non-compliance and management practice rather than system flaws. Incidents of neglect, abuse, and habitual

noncompliance are reported to unit managers and ACNO’s via e-mail, to facilitate individual coaching…



DMAIC

Feb 2006

What might be happening?

• If folks know what they’re supposed to do & how, but aren’t doing it, why?

– Simply forget/make mistake (why CMU tries multiple times)

– Can’t

• More pressing priorities (happens/understandable)

• Patient or nurse off-unit (preventable, with communication)

– Won’t (resistance due to misunderstanding or attitude)

• Don’t accept responsibility for taking appropriate action—”…not my problem…”

• Don’t appreciate potential risks

• Don’t understand accountability

• Don’t recognize that CMU isn’t aware of what’s happening on unit—only what the monitor says…

DMAIC





Feb 2006

Overall Conclusions

• Staff generally know what to do.

• Specifics of how maintenance is completed not clearly defined on every unit (when audited)

• Marked improvement in performance!

• Still some way to go toward goal…

• Still a gap between knowledge and behavior…

DMAIC

Feb 2006

2nd Remeasure (Alerts per Patient)

Remeasure Results - CMU Alerts per Monitored Patient

0.00

0.50

1.00

1.50

2.00

2.50

3.00

3.50

Leads 1.26 0.85 1.58

Other 0.27 0.36 0.72

Battery 0.74 0.47 0.62

Clinical 0.36 0.21 0.27

0-Before 1-After 2-After

Alerts/patient increased considerably—even exceeding baseline!

Most significant growth in lead-related alerts & in “Other”, which is

a non-specific “no signal” alert.

Alerts/patient increased considerably—even exceeding baseline!

Most significant growth in lead-related alerts & in “Other”, which is

a non-specific “no signal” alert.

Feb 2006

Drill-Down: Variability in Alerts per Patient

• Closer review of the data revealed special cause variation:

– 8CCP experienced an exceptionally high volume of SpO2-related calls on Thursday, Friday, and Saturday of the data period. This suggests at least one patient during this time was very non-compliant in wearing the SpO2 lead correctly, or at all! This is further evidenced by the high number of repeat alerts

– SpO2 usage was much higher toward the end of the week, driving up:

• Battery alerts (SpO2 drains batteries 3 times faster!)

• Lead alerts (which includes patients taking off the finger probe)

– Total patients on CMU telemetry does not vary significantly from day to day, but admissions & discharges both peak on Friday, increasingopportunities for errors. E.g., removing a transmitter to discharge a patient without first notifying CMU would drive up:

• Lead-related alerts

• “No signal”/other alerts

Feb 2006

Peak Alerts/PatientD

ata

DoW 7-Sun6-Sat5-Fri4-Thu3-Wed2-Tue1-Mon

5

4

3

2

1

0

Variable

NoSig/Pt

Battery/Pt

SpO2Lead/Pt

EKGLead/Pt

Types of Alerts per Patient, by Day of Week

Feb 2006

An unexpected new problem…

4/13/2005 1


















4/13/2005 1


















DMAIC

Some staff (willfully?) misinterpreted this to

mean “I never have to call CMU…”

Some staff (willfully?) misinterpreted this to

mean “I never have to call CMU…”

Feb 2006

Some Good News…

Code

AckR

t

2-After1-After0-Before

0.70

0.65

0.60

0.55

0.50

0.45

0.40

0.58

0.51

0.5750.581429

0.491429

0.558333

Boxplot of AckRt by Code

Code

Tim

eR

t

2-After1-After0-Before

0.50

0.45

0.40

0.35

0.30

0.25

0.20

0.37

0.27

0.215

0.337143

0.284286

0.223333

Boxplot of TimeRt by Code

Acknowledgement rate at 58%--new

high!

Acknowledgement rate at 58%--new

high!

Timely response at new high of 37%, vs. original of only 21%!

Timely response at new high of 37%, vs. original of only 21%!

Feb 2006

2nd Remeasure Summary

• Standardize battery policy, on every unit CMU supports:– Stock 9V batteries on the vitals checking devices.

– All CMU transmitter batteries to be changed when AM vitals check.

– For any patient with SpO2 monitoring, batteries must be changed when vitals checked on every shift.

• At discharge or when removing transmitter: CALL CMU first!

• Explore IS&T support in “automatically signaling” needs:– Battery & lead change requirements print on MAR or other nursing

patient care “checklists” (e.g., “pending orders”)?

– Discharges print to CMU?

• Emphasize to patients and families that wearing the SpO2 finger probe continuously & correctly is a responsibility in participating in their care/recovery.– Get orders to remove non-compliant patients from telemetry.

• Consider 48-hour automatic discontinuation to reduce opportunity for errors.

Feb 2006

Radically Rethinking “Central” Monitoring

• Human Nature– What gets measured…

– Out of sight = out of mind

– Everyone’s responsibility = nobody’s job

– Non-clinical (85%+ of alarms) = non-priority

• New experiment: What if responsibility were consolidated to one person per unit?– Advantages

• Local, & closer personal relationship to nurses

• Defined priority; doesn’t have to be a nurse!

• Less “noise” bothering nurses

– Disadvantages• Other tasks competing for attention

“If you always do what you’ve always done,

you’ll always get what you always got.”-

unknown

“Systems deliver exactly the results they

are designed to produce” – D. Berwick

“The significant problems we have

cannot be solved at the same level of thinking with which we created

them.” - Einstein

“If you always do what you’ve always done,

you’ll always get what you always got.”-

unknown

“Systems deliver exactly the results they

are designed to produce” – D. Berwick

“The significant problems we have

cannot be solved at the same level of thinking with which we created

them.” - Einstein

Feb 2006

The “Radical” Experiment

Concerns

• Competing priorities

• Disruptive to nursing unit

• Resistance from CMU

• Decentralizing Centralization?!!

Strategies

• Dedicated/only assignment

• “Watch” as well as “do”

• Volunteer unit

• Short pilot

• Emphasize benefits to RN’s

• Paint picture of success

• “Let’s try it and see…”

• Recognize “safety net”

• If it’s the wrong thing, we need to fix it!

Feb 2006

Results

• PCA identified & resolved 92% of “alert conditions” without having to be notified by CMU.

– Of those missed, the PCA was with another patient or at lunch.

– PCA addressed missed alerts upon notification

– Because PCA did not perform “preventive maintenance” in changing batteries & leads, volume of alerts was higher than necessary.

• Of all the alert conditions, 35 were clinical: 28* low SpO2 and 7 rhythm.

– Of these, the PCA checked the patient and resolved the issue before CMU called 31 of the 35 times (89%)

– Of CMU’s calls, 3 were low SpO2 & 1 was for heart rate

Results of "Radical" Experiment

* This number is somewhat higher than usual, as CMU often waits until a definite pattern of low SpO2 emerges, as patient behaviors can lead to false alarms or temporary drops that resolve themselves.

Feb 2006

Interpreting the Results

• More responsive: Consolidating responsibility to a single individual increased acknowledgement/resolution rate to 100%

• Faster response: Consolidating responsibility & watching the monitor increased timeliness of response to 100%

• Less “noise”: Watching the monitor reduced CMU call volumes 87% (using normalized SpO2 alert volumes—not the total the PCA acted on).

Results of "Radical" Experiment

Feb 2006

OpportunitiesResults of "Radical" Experiment

• Resistance from PCA’s & HUC’s

• Potentially higher cost

• Defeats goals of centrallizing monitoring?

• Higher response rate

• Faster response

• Minimal “maintenance alerts”

• Less “noise” for RN’s

• Sense of “local control”

• Assign “watching” & “doing” responsibility to an individual person, who could have minimal patient care training and who has basically no additional direct patient care load.

B

• Low utilization

• Increased training/competency costs

• Increased technology costs


• Faster response



• Sense of “local control”

• Decentralize monitoring back to nursing units

C

• Resistance from PCA’s & HUC’s

Disadvtgs/ Risks/

Barriers


• Faster response



• Improved accountability

• Minimal cost

Advantages

• DAY SHIFT ONLY

• Assign CMU maintenance & response accountability to one person (who could be a HUC!), who then carries a reduced direct pt care load.

• Augment CMU monitoring by having HUC watch monitors when not otherwise occupied.

Description

A

As CPOE implementation progresses, HUC role will change dramatically, but still be needed. This function has been

designated a priority as the new job description is developed and piloted.

As CPOE implementation progresses, HUC role will change dramatically, but still be needed. This function has been

designated a priority as the new job description is developed and piloted.

Feb 2006

Overview

• Defining Safety





• Summary

Feb 2006

Getting & Maintaining Support

• Mandated from the top

• Team-based: CMU + Nursing Units

• Regular progress reporting to unit managers

• Frequent outreach to nurses– Promotional

– Educational

– One-on-one

• Pilots with results reporting

• Process for ongoing monitoring & feedback

Feb 2006

Overview

• Defining Safety





• Summary

Feb 2006

Summary: The Actual Process

1. Define “safety”: what are the observable signs?

2. Determine metrics

3. Establish baseline performance

4. Document & analyze process– Flow charts

– Cause & Effect brainstorming (Fishbone)

– FMEA to prioritize

5. Identify solutions– Focused brainstorming on FMEA priorities

– Data-driven imperatives—”reap what you sow”

– Ensure behavior changes hold

6. Pilot & re-measure

7. Feedback mechanisms

Feb 2006

Questions?

Terry Dunn Six Sigma Black BeltHealth Alliance of Greater Cincinnati513/[email protected]

Ensuring Remote Telemetry Monitoring Patient Safety · Remote Telemetry Monitoring ... – Brain...

Documents

Transcript of Ensuring Remote Telemetry Monitoring Patient Safety · Remote Telemetry Monitoring ... – Brain...