Centre for Health Systems and Safety

Research

The impact of the electronic medical

record on hospital pathology

services – what is the evidence?

Associate Professor Andrew Georgiou

Senior Research Fellow

Outline • Background

o Existing evidence of the impact of

health IT on pathology services

o What do health professionals

think?

• Aim and Method

o What are the key performance

indicators?

• Results

o The impact on efficiency,

effectiveness and patient

outcomes and safety?

• Conclusion

Evidence of the impact of health IT

• 19 studies in total

• Most (7/11) showed

decreases in ordering

including a 27%

reduction in redundant

tests

• Four studies showed

improved adherence to

guidelines and improved

efficiency (up to 50%)

• Few studies across

multiple sites

Electronic ordering systems in the

Emergency Department

• Increase in time clinicians spend on

computers (up to 11% physicians

and 16% nurses)

• No change in time spent on patient

care

• Decreases in potential adverse drug

events (0.9/100 orders), laboratory

turnaround times

• 73% decrease in specimen

processing errors

What are health professionals

concerned about?

0%

5%

10%

15%

20%

25%

30%

35%

40%

45%

50%

secu

rity

inex

perie

nced

sta

ff

wor

k pr

actic

es

softw

are/

hard

war

e

educ

ation/

traini

ng

relatio

nshi

ps/com

mun

icat

ion

de-s

killin

gco

st

implem

enta

tion

roll o

ut

Issues of concern

Fre

qu

en

cy o

f p

art

icu

lar

issu

es a

s a

perc

en

tag

e o

f to

tal

co

ncern

s e

xp

ressed

by e

ach

pers

on

nel g

rou

p Senior staff - predominantly

management

Senior staff - management &

clinical

Staff - predominantly clinical

Research question

What is the impact of the

EMR on pathology

services, their work

processes and

relationships with other

departments, and on key

performance indicators?

Key performance metrics

Georgiou et al. Int J Med Info 2006

Test orderTest

processing

Test result

application

Costs Work practices

Test volumesRedundant test

rates

Guideline compliance

Turnaroundtimes

Doctor-lab communication

Patient management

Length of stay

Patient safety

Average turnaround time in minutes

Before implementation

(95% CI)

After implementation

(95% CI)

t test results*

All test assays 73.8 (72.2-95.4)

58.3 (57.1-59.4)

t=15.6 (df 184257)

p=0.000

Prioritised tests 44.6 (42.4-46.8)

40.1 (38.7-41.6)

t=3.3 (df 37830)

p=0.001

Non-prioritised

tests 81.5 (79.6-83.5)

65.9 (64.4-67.4)

t=12.6 (df 148493)

p=0.000

Tests in business

hours 81.8 (80.1-83.5)

69.0 (67.4-70.6)

t=10.7 (df 141219)

p=0.000

Tests outside

business hours 54.0 (50.6-57.4)

39.2 (37.8-40.5)

t=7.9 (df 37524)

p=0.000

Tests in control

ward 68.7 (63.9-73.5)

64.7 (60.4-69.0)

t=1.2 (df 12993)

p=0.218

Westbrook et al. (2006) J Clin Pathol

Test turnaround time significantly declined

Year 1 by 18.6% , Year 2 by 12.6%

Period No. of tests Mean in minutes

(95% CI)

All tests 2003 97851 35.35

(35.11,35.59)

2004 113752 28.77

(28.59,28.95)

2005 131022 25.14

(24.99,25.29)

TAT pre & post CPOE in four

hospitals

2005

Before 2006

After 2007

After Kruskal-Wallis

Hospital A - Median TAT 77 68 66 P<0.001

% tests using CPOE 75% 80%

Hospital B - Median TAT 145 129 108 P<0.001

% tests using CPOE 0-44% 57%

Hospital C- Median TAT 138 135 113 P<0.001

% tests using CPOE 29-38% 53%

Hospital D- Median TAT 141 139 128 P<0.001

% tests using CPOE 56-71% 74%

Median TAT in minutes

Volume of tests and

specimens* Average number of test assays per

patient did not change

92.5 assays/patient versus 103.2

(P=0.23)

Average number of specimens per patient

did not change

10.8/patient versus 11.7 (P=0.32)

*Westbrook et al. (2006) J Clin Pathol

Quality of pathology

ordering

Specification of

gentamycin specimens

Before 16% of gentamicin and 13% of vancomycin samples specified as peak or trough.

After significant increase - 73% for gentamicin and 77% for

vancomycin.

Westbrook et al. J Clin Pathol 2006

Heparin and warfarin

Proportion of patients reported to be on warfarin or heparin for aPTT & PT/INR

Pre-CPOE 3% (n=253) aPTT and 1.9% (n=161) PT/INR

Post-CPOE 3.9% (n=393) aPTT (p<0.005) and 2.6% (n=282) PT/INR (p<0.005)

Median laboratory turnaround time

Pre-CPOE 28 min. (aPTT) 34 min PT/INR Post-CPOE 21 min. (aPTT) (p<0.005) 23 minutes for

PT/INR (p<0.005)

The importance of quality

information exchange

• 43% of handwritten requests to

Microbiology contain patient related

clinical information e.g., “Recent

chemotherapy” “Previous MRSA”

“Recent travel”

• 97% of the clinical information were

deemed to impact on the

processing and/or interpretation of

lab tests

The impact of electronic ordering

on information exchange

Wound specimens with a request

specifying source and body site

Before electronic ordering (2005) 578 (69.6%)

One year later (2006) 774 (92.9%)

Two years later (2007) 814 (95.3%)

Three years later (2008) 877 (95.6%)

% of wound specimen requests where

admission reason was relevant

Before electronic ordering (2005) Not available

One year later (2006) 531 (63.7%)

Two years later (2007) 564 (66.0%)

Three years later (2008) 624 (68.0%)

Results from

regression analysis

Turnaround time was a significant factor

contributing to patients’ length of stay in

the emergency department. The model

accounted for 25.4% of variance (Adj.

R2=0.254)*

Laboratory errors

• Up to 68% of laboratory errors

occur in the pre-analytical test

ordering stage (Bonini, 2002)*

• Pre-analysis – all steps prior

to actual specimen analysis

• Errors can involve patient

and processing related

factors

*Errors in Laboratory Medicine Clin Chem 2002

Types of pre-analytical errors

• Mislabelled specimen

• Mismatched specimen

• Specimen suitability

• Leaking specimen

• Accident to specimen

• Insufficient specimen

9.7

12.7 13.5

0

4

8

12

16

2009 2010 2011

Error rate per 1000 test order episodes

N=1772

N=2282 N=2452

EMR errors = 464 EMR errors = 477

Errors per

1000 test order

episodes

Year

Incident Information Management System

(IIMS) reported errors

EMR Paper

Mislabelled specimen 0.1

(n=39)

0.31

(n=56) p<.001

Mismatched

specimen

0.49

(n=200)

1.42

(n=255) p<.001

Unlabelled specimen 1.37

(n=559)

1.65

(n=296) p<.01

“EMR test order problems”

Error details Total

Handwritten request on EMR 66.7% (n=451)

Order number problem (e.g.,

invalid number)

3.1% (n=21)

Multiple forms received

(Duplicates)

1.2% (n=8)

Incorrect order (e.g., swab

instead of fluid)

2.2% (n=15)

Change of tests 0.2% (n=1)

Add-on test 0.2% (n=1)

No information provided 26.5% (n=179)

Total 676

Comparison of Turnaround Time for test

orders with “EMR test order problem

EMR test order

problem

All test

orders

Median Data Entry time (mins) 8 5

Z=7.65, p<.001*

Median Total Lab TAT (mins) 263 82.14

Z=8.91, p<.001*

Total Episode Count (n=) 174 124119

*Wilcoxon signed-rank tests of significance

Missed test results

• Critical safety issue – increases

the risk of missed or delayed

diagnoses World Alliance for Patient Safety, WHO, 2008; Schiff, 2006

• Clinicians are concerned that their

test management practices are

not systematic Poon et al. Arch Int Med 2004

• Medico-legal concerns Berlin, AJR, 2009

• Impact on patient outcomes Roy et al. Ann Intern Med, 2005

How many results are missed for

hospital patients?

• Hospital inpatients 20% - 62% of tests are missed

• ED patients (discharged) 1% - 75% of tests are missed

Callen et al. BMJ Qual Saf 2011;20;194-199

• Ambulatory patients 7% - 62% laboratory tests missed

1% - 36% imaging tests missed

Callen et al. Jnl Gen Int Med, 2012

Study methods

Survey design (17 questions)

1 metropolitan ED; senior ED doctors

Significantly abnormal results

– not life threatening but need short-term

follow-up (e.g., chest x-ray with new shadow,

abnormal PSA)

Automatic patient notification methods

– Patient portal, Email, SMS, fax, mail or

phone

What types of tests were missed?

(%)

Are there standard policies and

procedures for patient notification of

results?

Who is responsible for notifying the

patient of a test result?

Perceptions of missed test results

19.2

26.9

53.9

In the past year I have missed an abnormal result that led to delayed

patient care

Yes (%)

No (%)

Don't know (%)

38.5

11.5

50

In the past year a colleague has missed an abnormal results that

led to delayed patient care

Yes (%)

No (%)

Don't know (%)

Safety considerations with health IT

implementation • Solutions need to be multipronged

• Policies, procedures and

responsibilities

• Role of patients, doctors, nurses,

clerical staff and laboratories in

the follow-up process

• Evaluation of information and

communication technology (ICT)

solutions

• Integrate solutions with work

practices of health professionals

Acknowledgements

Australian Research Council Linkage Grant (LP0347042) to evaluate

the impact of information and communication technologies on

organisational processes and outcomes: a multi-disciplinary, multi-

method approach (2003 – 2007)

Australian Research Council Linkage Grant (LP0989144) to

investigate the use of information and communication

technologies to support effective work practice innovation in the

health sector (2008 – 2012)

Quality Use of Pathology Program grant, Department of Health and

Ageing (2008-2009)

HREC Ethics approval Project No. 2005/058 and 2007/077

Thank you

Email: a.georgiou@unsw.edu.au

Website: www.aihi.unsw.edu.au

Twitter: @AGeorgiouUNSW