Using patient data for improvement work

Using patient data for improvement

work

Record reviews versus quality registries:What do we achieve with each method?

Kjersti MevikPhd student, MD

Patient safety advisor

Regional patient safety resource center Nordland Hospital trust

Bodø, Norway

13.04.16

Experiences from a Norwegian hospital trust

http://www.google.no/url?sa=i&rct=j&q=&esrc=s&frm=1&source=images&cd=&cad=rja&uact=8&ved=0CAcQjRxqFQoTCMKCk6bMo8cCFYEqcgod0SMM1g&url=http://www.dagbladet.no/2010/11/10/nyheter/nordlandssykehuset/feilbehandling/innenriks/helse/14237801/&ei=2ULLVYKyBIHVyAPRx7CwDQ&bvm=bv.99804247,d.bGg&psig=AFQjCNGvfbVPiL13IUkWF-pgYZu6XRoFNQ&ust=1439470525766693

http://www.google.no/url?sa=i&rct=j&q=&esrc=s&frm=1&source=images&cd=&cad=rja&uact=8&ved=0CAcQjRxqFQoTCMKCk6bMo8cCFYEqcgod0SMM1g&url=http://www.dagbladet.no/2010/11/10/nyheter/nordlandssykehuset/feilbehandling/innenriks/helse/14237801/&ei=2ULLVYKyBIHVyAPRx7CwDQ&bvm=bv.99804247,d.bGg&psig=AFQjCNGvfbVPiL13IUkWF-pgYZu6XRoFNQ&ust=1439470525766693

1. Fracture after patient falls in the bathroom

2. Wound infection after surgery

3. Blood transfusion after hip replacement

4. Urinary tract infections when using catheter

5. Gastroenteritis while admitted to hospital

Foreseeable complications

Adverse events

or

ERRORBLAME

HARM

SUFFERING

GTT definition of harm

”Unintended physical injury resulting from or contributed to by medical care that requiresadditional monitoring, treatment or hospitalization, or that results in death. ”

C1 Transfusion or use of blood products

C2 Code/arrest/rapid response team

C3 Acute dialysis

C4 Positive blood culture

C5 X-ray or Doppler studies for emboli or DVT

C6 Decrease of greater than 25% in hemoglobin or hematocrit

C7 Patient fall

C8 Pressure ulcers

C9 Readmission within 30 days

C10 Restraint use

C11 Healthcare-associated infection

C12 In-hospital stroke

C13 Transfer to higher level of care

C14 Any procedure complication

C15 Other

S1 Return to surgery

S2 Change in procedure

S3 Admission to intensive care post-op

S4 Intubation/reintubation/BiPap in PACU

S5 X-ray intra-op or in PACU

S6 Intra-op or post-op death

S7 Mechanical ventilation greater than 24 hours post-op

S8 Intra-op epinephrine, norepinephrine, naloxone, or romazicon

S9 Post-op troponin level greater than 40 ng/l

S10 Injury, repair, or removal of organ because of accidental injury

S11 Change in anesthesia procedure

S12 Insertion of artery catheter or central venous catheter

S13 Surgery more than 6 hours

S14 Any operative complication

M1 Clostridium difficile positive stool

M3 INR greater than 6

M4 Glucose less than 50 mg/dl

M5 Rising BUN or serum creatinine greater than 2 times baseline

M6 Vitamin K administration

M7 Benadryl (Diphenhydramine) use

M8 Romazicon (Flumazenil) use

M9 Naloxone (Narcan) use

M10 Anti-emetic use

M11 Over-sedation/hypotension

M12 Abrupt medication stop

M13 Other

I1 Pneumonia onset

I2 Readmission to intensive care

I3 In-unit procedure

I4 Intubation/reintubation

P1 Terbutaline use

P2 3rd- or 4th-degree lacerations

P3 Platelet count less than 50,000

P4 Estimated blood loss > 500 ml (vaginal) or > 1,000 ml (C-section)

P5 Specialty consult

P6 Oxytocic agents

P7 Instrumented delivery

P8 General anesthesia

P9 Apcar score <7 after 5 minute

P10 Induced labour

E1 Readmission to ED within 48 hours

E2 Time in ED greater than 6 hours

Triggers used in Global Trigger Tool

Categorization of adverse events

Types

Hospital acquired infections

Urinary tract infection

CVC infection

Ventilator associated pneumonia

Other infection

Lower respiratory infection

Surgical complications

Infection after surgery

Respiratory complications after surgery

Return to surgery

Injury, repair or removal of organ

Occurrence of any operative complication

Switch in surgery

Other

Bleeding/thrombosis

Thrombosis/Embolism

Bleeding

Bleeding after surgery

Patient fall /fracture

Patient fall

Fracture

Other

Allergy

Medical technical harm

Deterioration and chronic illness

Medication harm

Obstetric harm

Pressure ulcer

Severity

E- Temporary harm requiring intervention

F- Temporary harm requiring initial or prolonged hospitalization

G- Permanent patient harm

H- Intervention required to sustain life

I- Patient death

In our department all errors are responded to

Resistance to record reviews

The review method includes too few records

and we don’t have any more

resources to review more!

This harm did not happen to «our»

patients

The recordreview

method is not reliable!The record review

method does not findthe important adverse

events

What we did to overcome the resistance

• Involved the clincial staff

• The clinical staff analysed their «own» patients

• Developed an automatic search for triggers

• Adjusted the sample size

What is the correct sample size?

• Appropriate – do we review the relevant records?

• Correct for whom?

• Statistically correct?

Research is always good!

Evaluating the sample size:

Our sample size:70 records/bi-weekly

Recommended sample size: 10 records/bi-weekly

The rate differed

Mevik et al: Does increasing the size of bi-weekly samples of records influence results when using theGlobal Trigger Tool? An observational study of retrospective record reviews of two different sample sizes. BMJ Open 2016

0

10

20

30

40

50

60

70

80

90

100P

1 J

an 2

01

0

P2

Jan

20

10

P1

Feb

20

10

P2

Feb

20

10

P1

Mar

20

10

P2

Mar

20

10

P1

Ap

r 2

01

0

P2

Ap

r 2

01

0

P1

May

20

10

P2

May

20

10

P1

Ju

n 2

01

0

P2

Ju

n 2

01

0

P1

Ju

l 2

010

P2

Ju

l 2

010

P1

Au

g 2

010

P2

Au

g 2

010

P1

Sep

20

10

P2

Sep

20

10

P1

Oct

20

10

P2

Oct

20

10

P1

No

v 2

010

P2

No

v 2

010

P1

Dec

20

10

P2

Dec

20

10

Large sample size

Small sample size

Mean

Mean

But not the severity

0

10

20

30

40

50

60

70

E- Temporaryharm requiring

intervention

F- Temporaryharm requiring

initial orprolonged

hospitalization

G- Permanentpatient harm

H-Interventionrequired tosustain life

I- Patientdeath

Percentages

Large sample

Small sample

Or the types

0

5

10

15

20

25

30

35

40

45

Pe

rce

nta

ges

Large sample size

Small sample size

Implications of the study

To evalute the rate of adverse events:

• Adjusting sample size to hospital size?

• An increased new fixed sample size?

To evalute the types and severity of adverseevents:

• 10 records reviewed bi-weekly is sufficient

How are we using the results from theGTT?

If you don’t use them it is not necessary to measure!

1.Sharing the results

1.Transparency

2.Visible and available for involved staff

3.Simple and clear presentation

4.Updated data – fresh data dont smell

5.Encounters – the more the merrier

Transparent and visible – also for patients

2. Tracking the rate of adverse events over time

3. Evaluating the effects of different initatives

0

2

4

6

8

10

12

14

16

2010 2011 2012 2013 2014

Number of postoperative bleedings

0

5

10

15

20

25

30

35

2011 2012 2013 2014

Number of postoperative infections

0

10

20

30

40

50

2010 2011 2012 2013 2014

Number of adverse drug events

4. GTT results compared to other methods

Voluntary incident reporting:

How many percent of the adverse events identified in the GTT was also reported by the voluntary incident reporting system:

0-1 % 10-40 % 80 -100 %

5. Getting the boards on board

Record reviews

• Identifies types and severity of adverse events

• Tracking rate of adverse events over time

• Evaluate the effects of improvement work

• Understandable results

• All cause harm- includes the whole patient population

Meetbaar Beter

Dr. Jan-Henk Dambrink, MD

Member Advisory Board Meetbaar Beter

Interventional Cardiologist

Isala Zwolle, The Netherlands

Dr. Paul van der Nat

Manager Development & Implementation

Meetbaar Beter

Meetbaar Beter | International Forum on Quality & Safety in Healthcare 2016

“You will hear how Meetbaar Beter, an

initiative of 19 heart centers, uses outcomes

measures to realize improvements in care

delivery – even for medical conditions with

low patient volumes.”

26


The Meetbaar Beter centers

Participating heart centers

Participating intervention centers


Part 1: Introducing Meetbaar Beter

(Paul van der Nat)

Part 2: Physicians perspective: how

Meetbaar Beter has enabled us to improve (Jan-Henk Dambrink)

Content

28


Statement 1

A quality registry

without

connected improvement initiatives

has little value


Aim of Meetbaar Beter

By using a solid scientific basis, Meetbaar Beter defines and

measures relevant outcomes of cardiovascular treatments

to improve transparency and quality of healthcare delivery

for heart patients.

30

#1 Transparent outcome reporting

Showing all heart centers no matter what the results….

31


Covering more than 90% of Dutch heart centers

32


For the four main medical conditions including all treatment options (99% completeness in 2015).

33


#2 not (just) a registry

Outcomes are used as starting point for improvements

34

Year

Nu

mb

er

2010 2011 2012 2013 2014 2015 2016

0

2

4

6

8

10

12

14 Improvement initiativesIndicator sets



Outcome reports are used to create a learning & sharing platform

35



Leading to the identification and roll-out of potential best practices.

36

#3 physicians in the lead

Medical leadership arranged within: board of directors, advisory board, Outcomes Teams (development), Medical experts (data audits).

37


#4 Complete patient populations

Outcome measures sets are defined for medical condition, including all treatment options, thus avoided unwanted risk selection.

38

Coronary Artery Disease

Treatment independent outcomes

CABG

Outcomes related to treatment

PCI


Conservative



#5 Patients at the center

Outcomes are presented for patients using infographics

39


#5 Patients at the center

Outcomes are selected within the Hierarchy of Porter including clinical outcomes as well as PROMs.

40


○ Checks on peculiar outcomes and/or statistical outliers.

○ Annual audits on process and outcome data by medical experts

#6 Building excellent data quality

Several quality controls are performed to ensure data quality: trend analyses, logic analyses, and audits in the participating centers

41


#7 Continuous innovation

Meetbaar Beter has defined the outcomes-based improvement cycle and supports realization in each of the heart centers.

42

3. Selection improvement

initiatives

1. Monitoring of

outcomes

4. Implementation of

improvements

2. Identification

improvement potential

a. Development outcome

measures sets

b. Maintenance of sets based on

physician and patient use of data

c. Data quality audits

d. Outcome reporting

a. Identify internal improvement potential

b. Identfiy external learning potential

c. Data mining, additional statistical analyses

a. Internal improvement cycles

b. Sharing best practices

c. Platform for learning&sharing

a. Roll-out of best practices

b. Monitoring of progress

c. Scientific evaluation of best

practices


Content

43

Part 1: Introducing Meetbaar Beter

(Paul van der Nat)

Part 2: Physicians perspective: how

Meetbaar Beter has enabled us to improve (Jan-Henk Dambrink)

44

Sweden is a quality leader in Cardiology

http://www.ucr.uu.se/swedeheart/

European CARDS initiative - 2004

Eur Heart J. 2005 Feb;26(3):308-13

European CARDS initiative - PCI

○ CARDS included 87 variables on the patient and the procedure, and a further 25 on outcomes and follow-up

○ Meetbaar beter: only 12 initial conditions and 6 outcome indicatiors!

○ Initial conditions and outcome indicators carefully selected by an International Advisory Council

| Title of the presentation | 01-01-2013 page 46

○ Result: near complete dataset (published data > 90% complete)

○ Clinically relevant outcomes

○ Reliable statistical models. • Example: predictive power 30-day mortality after PCI

C-statistic 0.89.

○ Detect improvement potential – reveal ‘champions’ in the field

Meetbaar Beter

47


Statement 2

Quality improvements

should only be based

on

statistically significant data


Presentation of data:

allowing maximum learning potential

49

(1) Raw outcomes

(2) Segmented figures (3) Risk-corrected funnelplots


○ Mortality after PCI in kidney failure

○ Vascular complications in TAVI

○ Isala Safety Check

Examples of improvement projects

50


Segmented figures: Mortality PCI for

patients with kidney failure.

51

30-days mortality after PCI - 2013

1.4% 9.2% 1.9%

0%

2%

4%

6%

8%

10%

12%

14%

51/3629

no kidney insufficiency

21/228

kidney insufficiency

0/6

unknown

72/3863

overall

Relative Risk 6.6 (3.9 – 10.9)

2.0% 6.3% 3.2%

0%

2%

4%

6%

8%

10%

12%

14%

89/4363 57/910 26/92 172/5365

Relative Risk 3.1 (2.2 – 4.3)

OLVG vs Catharina p=0.115

Amsterdam

Eindhoven

52

○ Stop diabetes medication 48 hours before PCI

○ Give additional fluids before and after procedure

○ Limit contrast agent to max. 300 ml

○ No additional PCI procedures < 1 week

Standard measures already in place (before 2013):

Mortality after PCI in patients with kidney

failure: improvement program


53

○ No planning of PCI procedures without recent knowledge

of kidney function.

○ Differentiation in severity of kidney failure with regards to additional fluids

○ Weight related formula for max. amount of contrast agent

○ No additional PCI’s < 4 weeks

○ Encouragement to use smaller catheters, leading to less contrast use.

Mortality after PCI in patients with kidney

failure: improvement program

Additional measures (after 2013):


54

9.20%

5.00%

0%

2%

4%

6%

8%

10%

Voor Na

30-days mortality for patients

with kidney failure

Segmented figures: Mortality PCI for


Before After

Absolute mortality reduction of – 4.2%

55

Segmented figures (2): Mortality PCI for


30-days mortality after PCI - 2015

0%

2%

4%

6%

8%

10%

12%

14%

AMC

(n=5387)

Amphia

(n=0)

Antonius

(n=5517)

Catharina

(n=8589)

Erasmus

(n=1229)

Haga

(n=3237)

Isala

(n=6400)

MCL

(n=4259)

MST

(n=5092)

MUMC

(n=0)

OLVG

(n=4411)

Radboud

(n=1498)

UMCG

(n=0)

UMCU

(n=2141)

Coronary artery disease | PCI | 30-days mortality | kidney insufficiency

no kidney insufficiency kidney insufficiency© Stichting Meetbaar Beter

OLVG: Relative Risk kidney failure < 2





56

57

Transfemoral

Transcutaneous Aortic Valve

Implantation

58

Vascular complications after TAVI

Amphia Breda: 90% transfemoral


59

Risk-corrected funnelplot: Vascular

complications after TAVI

Amphia Hospital: ‘access champions’:

- Surgical access in stead of puncture

- Preference for small diameter systems





60


Isala safety check

61

Significantly reduced mortality after bypass surgery:

What was their secret?


62

Isala Safety Check in 5 rules (2 min)

1. Time out

2. EuroSCORE: Risk stratification

3. Assess cardiac, pulmonary, renal, neurological risk; infections, bleeding risk

4. Echocardiographic findings (transoesophageal)

5. Adjust treatment plan if indicated

Veiligheidscheck cardiochirurgie versie 8-14032013

VEILIGHEIDSCHECK

Regelvan5TeamApproach

Cardiochirurgie……………….

1 VOORINLEIDING:……………….SAFETYITEMS“ALGEMEEN”………………………………………………

TimeOutproceduregevolgd Ja Nee

2 VOORINCISIE…………………….Safetyitems“SPECIFIEK”…………………………………………………

EuroSCORE(Standaard) Laag0-2 Gemiddeld3-5 Hoog≥6

3 ….…VERHOOGDRISICOt.a.v.:………. Nee Ja Bespreek

1.Cardiaal2.Pulmonaal3.Renaal4.Neurologisch5.Infectieus6.Bloedingsstatus

4 …………Safetyitems“TEE”……………… Normaal Afwijkend Bespreek

1.LV-Functie/LAA/OFO2.Kleppen

3.Atheroscleroseascendensetc.

5 ……AANPASSINGPROCEDURE….. Nee Ja Bespreek1-8hieronder

1.IABPinbrengen:chirurgisch/percutaan2.Epi-aorticscanningnaTEE3.Geenchirurgie:onvoldoende-geenindicatie/ernstigverkalkteaortaasc./“in-operabel”4.Klem:andereplaatsaorta5Canulatie:andereplaatsopaortaascendens/andereplaatselders/anderecanule6.Pomp:oppomp-zonderAOXklem/off-pump7.Intracardialeaanpasssing:+/-MVP,+/-AVR,+/-TVP,LV-LAAthrombus,+OFO8.Proximaleanastomose:geensidebitingklem/andereplaatsaortaasc/LIMA-T/Ygraft9.Extracardialeaanpassing:patchin–ofvervangingaortaascendens10.Anders:beschrijf

….TEAM:CARDIOTHORACICSURGEONS-CARDIOTHORACICANESTHESIOLOGISTS….

BBB RL TM GS HS SS WS Other

EE JE CK MK BM AN CR WS FS HS SSP Other

Patiëntsticker

What is the Isala safety check?

2010N=1311

2011N=1270

2012N=1240

Expected

mortality

Actual

mortality

Expected

mortality

Actual

mortality

Expected

mortality

Actual

mortality

CABG 5,5% 1,0% 5,2% 0,6% 4,5% 0,2%

AVR 8,8% 2,4% 6,7% 0,8% 8,1% 0,7%

CABG + AVR 9,7% 3,4% 7,7% 1,0% 9,6% 2,0%

Other surgery 13,8% 5,0% 14,1% 5,8% 12,4% 4,9%

Overall 8,5% 2,5% 8,1% 2% 7,7% 1,8%

Overall in hospital mortality decreased each year with 15%

63

Mortality after surgery in Isala

64

Impact Isala safety check

○ Meetbaar Beter has developed a

comprehensive and near complete database of

patients with heart disease that receive invasive

treatment in The Netherlands

○ This database allows for benchmarking between

centres and for a comparison with data from the

literature using weighted results

Conclusions (1)

65


○ The examples shown demonstrate that:

• Even non-significant differences using segmented results can help to identify improvement potential

• a study of outliers in risk-corrected funnelplots can help to identify champions in the field or improvement potential in individual centres

○ Sharing these results can help all centres to improve their patient outcomes.

Conclusions (2)

66


Thank you for your

attention!

Thank you!

68

[email protected]

[email protected]


Statement 1

A quality registry

without

connected improvement initiatives

has little value


Statement 2

Quality improvements

should only be based

on

statistically significant data


Using patient data for improvement work

Documents

Transcript of Using patient data for improvement work