testing process error rates in a ugandan laboratory

8/7/2019 Testing Process Error Rates in a Ugandan Laboratory

1/23

Testing Process

Error Rates in a

Ugandan Laboratory

Karugaba Patrick1, Elbireer Ali2(1QA/QC Coordinator, MU-JHU Core Lab, 2Johns Hopkins Univ. Baltimore, MD,USA)

MU-JHU Core Laboratory,

Kampala Uganda


2/23

MU-JHU Lab Location at IDI

Building, Mulago Hospital


3/23

Introduction

Makerere University Johns Hopkins University

(MU-JHU) Core clinical & research lab is located at

the Infectious Diseases Institute (IDI), New Mulago

Hospital, Kampala.

Began operation in 1989, acquired CAP

accreditation 2003, fully complies to GCLP

standards & US CLIA-88 requirements Supported lab evaluations for over 60 clinical trials

since inception 20 years ago

Handles approx. 16,000 test requests monthly


4/23

Some past achievements/ recognitions

MTN best PT performance, May 2008

MLO Lab of the year 2nd runner up, April 2008 Zero deficienciesCAP inspection, 2009

MLO Lab of the year 1st runner up, April 2010


5/23

Prevalence of errors in

perspective

ISO defines lab error as any defect from orderingtests to reporting results and appropriateinterpretation and reacting on these

US IOM estimated that 44,000-98,000 Americansdie not from medical conditions they checked inwith, but preventable medical errors, Khon &Donaldson, 2000

This makes medical errors the eighth leadingcause of death in America no statistics availablefor other countries

World lab error rates vary greatly (0.1 9.3)%


6/23

Errors could likely occur at:-

Pre-analytic stage

Appropriate test

selection, ordering

Requisition form errors

Poor patient preparation

Sample collection

Specimen

transportation

Specimen reception &

data accessioning

Specimen processing

Analytic stage

Methods validation

Calibration & QC testing

Instrument maintenance

& trouble shooting

Testing procedure

performance

Results interpretation

Post Analytic stage

Results data entry

Results reporting


7/23

Core Lab error tracking/follow

up processes/mechanisms

SOP in place for all procedures, regularly reviewed

Unsuitable specimens tracking (hemolyzed, clotted, pooridentification, lipemic, insufficient quantity)

Data entry 2nd check Daily, weekly, monthly, semi-annual reviews of tests

QC, results review, instrument/equipment maintenance & functionchecks, comparability, competency assessments, team meetings,results interface accuracy check)

Results QC review prior to release to clinics Process improvement report

Customer service concerns (CSC)

Amended reports


8/23

2009 Error tracking

Un-acceptable/ rejected specimens 0.38%

Amended reports; 0.13% of all 2009 testing activities

86% were due to demographic errors (43.9% - lab,42.1% - Clinic). Clinic & lab appeared to have nearlyequal error contribution.

1.3% due to analytical errors, 12.7% others.

Internal - Process improvement reports (PIR)0.06% of testing activities

External Customer service concerns 0.04% oftesting activities.


9/23

Objectives

To determine the root causes of lab errors

leading to customer service complaints

To categorize identified errors as pre-analytic, analytic and post-analytic


10/23

Methods

Core lab instituted a mechanism to receivefeedback from clients regarding quality of servicesthrough the CSC form.

Also, when lab notes any errors of clinicalsignificance after results have been released,clients are requested to fill in a CSC form

Review was made of customer service complaintsreceived at the core lab for the period from Jan.

Dec. 2009 Root causes of problems were identified and

categorized as pre-analytic, analytic or postanalytic.


11/23

MethodsTools employed

The Customer service concern (CSC) form

was employed for data collection.

The CSC captured information on details ofspecimen(s) queried, details of

problem/concern and detail of lab response

including appropriate categorization of error.

Monthly analysis was made for CSC and a

report was compiled at the end of the year.


12/23

Results CSC Reports

analysisSUMMARY OF EXTERNAL QUALITY ASSURANCE REPORTS (CSC) FOR 2009

Category Item Total

% of

Cat Cat. Total

Per% of

Total

Pre-Analytical

Delay in Specimen pick-up/Processing 0 0.0

53 77%

Data Entry Error 35 66.0

Specimen Processing Error 7 13.2

Unacceptable/Rejected Specimen not called to Clinic 0 0.0

Pre-Other 11 20.8

Analytical

Wrong Tests Performed 0 0.0

6 9%Results Error Suspected 4 4.0

Tests Not Performed 1 16.7

Analytical-Other 1 16.7

Post - Analytical

Critical results not Called 4 40.0

10 14%

Results not delivered within expected TAT 0 0.0

Results Transcription Errors 1 10.0

Data Download Error 0 0.0

Yellow Form missing 0 0.0

Storage Status error 4 40.0

Post-Other 1 10.0

Total Per Month 69


13/23

esu s: ummary o a

errors as per CSC received -

2009SUMMARY OF EXTERNAL QUALITY ASSURANCE REPORTS (CSC) FOR 2009

05

1015202530

3540

Delay

inSp

ecimen...

DataEn

tryError

Speci

menP

rocessi.

.

Unacc

eptab

le/Reje.

..

Pre-Othe

r

Wrong

Tests

Perfo...

Resul

tsErr

orSusp..

.

Tests

NotP

erformed

Analy

tical-Oth

er

Critic

alresu

ltsnot...

Resul

tsno

tdeliver

..

Resul

tsTra

nscripti..

DataDo

wnloadEr

ror

Yellow

Form

missing

Storag

eStatuserr

or

Post-Othe

r

rror ype

oofco

firmedco

mplaits


14/23

Resul

ts:Pre-analyti

cerr

(CSC)

Su

aryo

Pre-

nalytical

rrors(C

SC)

0

3

5

7

0

0.0

66.0

13.2

0.0

010203040506070

Delay in

Specimen

pick-

up/Processing

Data EntryError

Specimen

Processing

Error

Unacceptable/

Rejected

Specimen not

called to Clinic

rrorcate

ory

o o errors


15/23

Results: Categorization of

errors

Categorization of errors lea ing to CSC

53

610

77

914

0

10

20

30

40

50

6070

80

90

P

A A

P

A

Errors Categor

Value

P


16/23

Discussion: Core Lab Error

rate

Data entry errors alone contributed to 50,1% of allCSC reports Total pre-analytic errors were 77%(0.03% of all testing activities)

For all errors of clinical significance, a clients wererequested to write a CSC.

Core lab total error rate recorded for 2009 by followup of CSC therefore was 0.04% of all testingactivities done in that year (4 of every 10,000

tests). Other error sources not covered by CSC analysis

such as sample quality, patient variables (diet,exercise, e.t.c.) not considered in this report.


17/23

Error rates from elsewhere

over the times

Historical lab error rates

1 error for

every % Error rate Year Source

30 - 50 2.5 McSwiney and Woodrow 1969

Guiseppe,

et al.2004

50 - 100 1.5 Sourverjin et.al. 1980

330 0.30 Chambers et. Al 1986

1000 0.10 Boone 1990

8300 0.01 Lapworth and Teal 1994900 0.11 Tutting et. Al. 1996

214 0.47 Plebani and Carraro 1997

164 0.61 Stehl et. Al. 1998

283 0.35 Hofgartner and Tait 1999


18/23

More recent error rates

Naadira et al, March 2010 reported 0.151% errorrate due to registration/ data entry errors alone atTygerberg Hospital, Cape Town, South Africa.

Reported error rates in-patient vs Outpatientservice are 0.6% & 0.039% respectively

Suggestive that patient care involving non-labpersonnel accounts for majority errors.

Bonini et al., 2002 reported literature reviews of

overall lab error rates giving Pre-analytical (31.6-75)%

Analytical (13,3-31.6)%

Post analytical (9-30.8)%


19/23

Causes of varying error rates

Poor or insensitive error detection

mechanisms

Under reporting of errors No universally accepted error definition

Extreme variation (heterogeneity) in

designs/strategies employed in lab errorstudies.

Different ways of classifying errors


20/23

Discussion

Elissa, 2004 urged that people have a great

role in quality laboratory systems

David A. Novis, Sep. 2008 agued:- Are wenot smart enough? Are we lazy? Do we just

not care?...It is not the people, it is the

system

Benchmarking:- define measurable quality

indicators, collect data, perform statistical

analysis, evaluate performance/outcomes.


21/23

Current extra Initiatives taken

at the core lab

Currently collecting information daily on data

entry error types and frequency

Initiated a six-sigma lean study in theprocessing department to help identify root

causes and mechanisms for improvement.

Opportunities for automation to reduce dataentry errors e.g. use of bar coding still not

feasible & processing equipment.


22/23

Conclusion

Error rates at Pre-analytic phase of the

testing process account for highest fraction

of lab errors at the core lab and similartrends are continually reported worldwide.

Data entry, patient information registration

and sample quality errors still account for

majority pre-analytic problems.

Focus should continue on staff training and

most importantly on improving systems.


23/23

Thank you

Thank you

testing process error rates in a ugandan laboratory

Documents