Content uploaded by Rixin Jamtsho
Author content
All content in this area was uploaded by Rixin Jamtsho on May 01, 2014
Content may be subject to copyright.
Content uploaded by Rixin Jamtsho
Author content
All content in this area was uploaded by Rixin Jamtsho on May 01, 2014
Content may be subject to copyright.
EXTERNAL QUALITY ASSESSMENT SCHEME IN CLINICAL
CHEMISTRY FOR DISTRICT LABORATORIES IN BHUTAN
(EQAC-DLB)
RIXIN JAMTSHO
A THESIS SUBMITTED IN PARTIAL FULFILMENT
OF THE REQUUIREMENTS FOR
THE DEGREE OF MASTERS OF SCIENCE
(MEDICAL TECHNOLOGY)
FACULTY OF GRADUATE STUDIES
MAHIDOL UNIVERSITY 2010
ACKNOWLEDGEMENTS
The achievement of the objectives in this project is highly attributed to an
intellectual and valuable guidance of advisors and many helping hands of my
colleagues. Therefore, I would like to thank Asst. Prof. Dr. Wilairat Nuchpramool, the
major advisor, Assoc. Prof. Phannee Pidetcha and Asst. Prof. Dr. Ratana Lawung, the
co-advisors for their valuable guidance and continuous support during the course of
my project.
My sincere gratitude also goes to Mr. Kamol Kotivongsa, Mr. Lerson
Suwungton and Ms. Umaporn Pinyosirikul, Department of Clinical Chemistry at
Siriraj Campus for arranging the equipments and reagents for this project; and also
staff of the Centre of Medical Laboratory Services, Faculty of Medical Technology for
testing my samples.
I have the special thanks for Mr. Dorji (Microbiologist), the Head of the
Department of the Laboratory Services at the Jigme Dorji Wangchuk National
Referral Hospital in Thimphu who was one of the pillars of the success in this work.
I am grateful to the laboratory staff and the staff of the administration at
the district hospitals in Bhutan for their support and our laboratory colleagues and
Bhutanese friends studying in Thailand who donated their blood for this project.
I would also like to thank all the staff of the Faculty of Graduate Studies
and the Faculty of Medical Technology, Mahidol University for their care and support
during my studies in Thailand and the Ministry of Health, Royal Government of
Bhutan for continuous financial support in my studies.
Finally, I‟m grateful to my beloved family for their advices and for
providing me a very peaceful and comfortable atmosphere during the course of my
studies in Thailand.
EXTERNAL QUALITY ASSESSMENT SCHEMES IN CLINICAL CHEMISTRY
FOR DISTRICT LABORATORIES IN BHUTAN (EQAC-DLB)
Abstract
External Quality Assessment Scheme (EQAS) involves the evaluation of a
performance of laboratories by an outside agency. The EQAS organizer as the outside
agency supplies a Quality Control (QC) material to be analyzed by the participating
laboratories whose performance is evaluated and monitored. However, many
developing countries face various limitations due to which EQAS has not been
established. The aim of this research was to prepare a home-made lyophilized human
serum QC material for the implementation of EQAS in Bhutan.
A preliminary study was conducted for this research through
questionnaires. Lyophilized human serum was prepared and distributed to 19
laboratories performing clinical chemistry tests. Nine routine analytes (glucose, BUN,
creatinine, AST, ALT, ALP, total bilirubin, total protein, and albumin) were included
in the study. Their results were statistically analyzed and performance was evaluated
using overall CV, variance index scores (VIS), and modified capability index (mCpk).
After the implementation of EQAC-DLB, CVs were gradually decreased
which indicated the improvement in the overall performance. The CVs which were
much higher in the beginning moved closer to the CLIA proficiency testing (PT)
criteria at the end of the study. The percentage results in acceptable VIS were
increased from 63% in November 2009 to 79% in July and 68% in August, 2010. The
maximum mCpk score was 165 and the minimum was -172. The gradual decrease in
the CVs and VISs showed improvement in the performance through our feedback.
We concluded that, the establishment of EQAS by preparing home-made
lyophilized QC material will be a useful scheme to monitor a laboratory‟s
performance in Bhutan and mCpk could be used as an alternative scoring index which
reflects both bias and precision.
Key words: EQAC-DLB / Variance Index scores / homemade; Lyophilized control
material / modified; Capability index.
CONTENTS
Page
ACKNOWLEDGMENTS
iii
ABSTRACT
iv
LIST OF TABLES
ix
LIST OF FIGURES
x
LIST OF ABBREVIATION
xii
CHAPTER
I
INTRODUCTION
1.1
A brief history of health services in Bhutan
1
1.2
Laboratory services in Bhutan
2
CHAPTER II
OBJECTIVES
4
CHAPTER III
LITERATURE REVIEW
3.1
Quality assurance in clinical laboratory
5
3.1.1 Internal quality control
5
3.1.2 EQAS
7
3.2
Quality control material
8
3.2.1 The matrix effect on QC material
9
3.2.2 Types of QC material
10
3.2.2.1 Frozen QC material
10
3.2.2.2 Stabilized liquid control material
10
3.2.2.3 Lyophilized QC material
10
3.3
Tests for sufficient homogeneity
11
3.4
Tests for sufficient stability
12
3.5
Evaluation of EQAS Results
12
3.6
Assigned values for EQAS
13
3.7
Limit of acceptability
14
3.8
The performance scores
14
CONTENTS (cont.)
Page
CHAPTER IV MATERIALS AND METHODS
4.1General materials
17
4.1.1 Equipments
17
4.1.2 Chemicals & Reagents
18
4.1.3 Miscellaneous items
18
4.2 Methods
4.2.1 Study on the feasibility of implementing EQAS
18
4.2.1.1 Collection of information on health systems
18
4.2.1.2 Collection of information using questionnaire
19
4.2.1.3 Assessment of basic QA knowledge
20
4.2.2 Preparation of control material
4.2.2.1 Collection of specimen
20
4.2.2.2 Specimen processing
20
4.2.2.3 Label adjustment
21
4.2.3 Study on the characteristics of the control material
4.2.3.1 Homogeneity testing
23
4.2.3.2 Establishment of assigned values
24
4.2.3.3 Stability of control material
24
4.2.4 Implementation of EQAS in Bhutan
4.2.4.1 Allocation of the samples
24
4.2.4.2 Result recording forms and package inserts
25
4.2.4.3 Design of the scheme for monthly analysis of
QC sample
25
CONTENTS (cont.)
Page
4.2.4.4 Distribution of QC sample to the participants
26
4.2.4.5. Performance monitoring and feedback reports
27
4.2.5 Performance assessment
4.2.5.1 Assessment of interlaboratory precision by
comparing monthly CV with CCV from CLIA PT
28
4.2.5.2 Evaluation of performance using VIS system
29
4.2.5.3 Evaluation of performance using mCpk
29
4.2.5.4 Assessment of the EQA in relation to the IQC
31
CHAPTER V RESULTS
5.1 Results from the study on the feasibility of implementing EQAS
5.1.1 Information on health systems using the official websites
33
5.1.2 Information from the questionnaires
34
5.1.3 Laboratory staff and qualification
36
5.1.4 IQC and EQA activities
37
5.2 Result of the study on characteristics of QC sample
5.1.1 Homogeneity test results for the EQA samples
38
5.1.2 Assigned values derived from the results of
ISO 15189 certified laboratories
39
5.1.3 Results of the stability tests
40
5.3 Results of the assessment of interlaboratory precision
45
5.4 Performance assessment using MVIS
5.4.1 Assessment on instrument and method performance
46
5.4.2 Assessment of the performance on 4concentration levels of
QC material
51
CONTENTS (cont.)
Page
5.4.3 Laboratory performance categorized based on VIS
52
5.5 Modified Cpk as performance indicator
54
5.6 EQAS performance in relation to the IQC
56
5.7 Analysis of performance characteristics of laboratory with highest
OMVIS in EQAS and high imprecision in IQC
57
CHAPTER VI
DISCUSSION
59
CHAPTER VII
CONCLUSION
66
REFERENCES
67
APPENDICES
74
Appendix A
Questionnaire
75
Appendix B
Result recording form
80
Appendix C
Feedback report form
81
Appendix D
Package inserts
83
BIOGRAPHY
85
LIST OF TABLES
Table
page
3.1
Summary of the eight international EQAS organizations
8
4.1
Level adjustments: adding selected biological compounds
23
4.2
Design of the EQAS rounds with trial numbers
27
4.3
Schedule for sample analysis and reporting for EQAC-DLB
29
4.4
Interpretation of the modified capability index (mCpk)
32
5.1
Current Health Centres in Bhutan
34
5.2 The bed size of hospital and number of laboratories in Bhutan using
various clinical chemistry analyzers
35
5.3 Methods/principle and volume of sample required per run for the
laboratories in Bhutan
36
5.4
The manufacturer‟s range used as IQC range by laboratories in Bhutan
36
5.5
Category of staff in the laboratories
37
5.6
IQC and EQA activities in the clinical chemistry laboratory in Bhutan
38
5.7
Results of the homogeneity test for the EQAC-DLB samples
40
5.8 The mean of the results from three ISO 15189 certified laboratories
designed as assigned values of control material in this project
41
5.9
Results of the stability test for the sample level A
1
stored at 2-8
o
C
41
5.10
Results of the stability test for the sample A
1
stored at -20
o
C
42
5.11
Stability results of EQA sample N
1
stored at 2-8
o
C
42
5.12
Stability results of EQA sample N1 stored at -20
o
C
43
LIST OF FIGURES
Figure page
4.1 Map of Bhutan showing the location of hospitals and laboratories
in Bhutan
20
4.2 Work flow for preparing homemade lyophilized human serum for
implementing EQAC-DLB
24
4.3 Flow diagrams for the sample distribution from Thailand to Bhutan
and to the participants of EQAC-DLB
28
4.4 Overall work plan for the activities carried out to implement
EQAC-DLB
33
5.1The pre-test and post-test scores of the 27 laboratory staff.
39
5.2 The distribution of the mean of the duplicate results of sample level
A
1
stored at 2-8
o
C and -20
o
C analyzed in 3, 5, 7, 8 and 9 months
44
5.3.
The distribution of the mean of the duplicate results of sample
level N
1
stored at 2-8
o
C and -20
o
C analyzed in 3, 5, 7, 8
and 9 months
45
5.4. Average CV of the results of two trials in the beginning and at the
end of the program compared with CCV from CLIA PT Criteria.
47
5.5
Methods and instruments performance on glucose
48
5.6
Methods and instruments performance on BUN
49
5.7
Methods and instruments performance on creatinine
49
5.8
Methods and instruments performance on AST
50
5.9
Methods and instruments performance on ALT
50
5.10
Methods and instruments performance on ALP
51
5.11
Methods and instruments performance on total bilirubin
51
5.12
Methods and instruments performance on total protein
52
LIST OF FIGURES (cont.)
Figure
Page
5.13 Methods and instruments performance on albumin
52
5.14 Mean VIS of each analytes were compared among four levels
of the control material. A1 and A2
53
5.15
The performance evaluation by grouping VIS into three categories
54
5.16
OMVIS for each month of the study
55
5.17
mCpk for glucose among 19 laboratories
56
5.18 Cpk for glucose among 19 laboratories 56
5.19 MVIS of glucose for each laboratory
57
5.20 OMVIS for each laboratory in Bhutan. The Lab-1418 which shows
the highest OMVIS was chosen to monitor the performance during
the entire period of study.
58
5.21 The MVIS of glucose for the Lab-1418 during the entire period
of implementing EQAC-DLB.
59
5.22 The MVIS of AST for the lab 1418 during the entire period
of implementing EQAC-DLB
. 60
LIST OF ABBREVIATIONS
ACBI-CMC Association of Clinical Biochemistry of India,
Christian Medical College
ALP Alkaline phosphatase
ALT Alanine aminotransferase
ANOVA Analysis of variance
AON Average of normal
AV Assigned values
AST Aspartate Aminotransferase
BCG Bromocresol Green
BHU Basic Health Unit
BUN Blood urea nitrogen
BV Biological variation
o
C Degree Celsius
CAP College of American Pathologists
CCV Chosen coefficient of variation
CDC Center for Disease Control
CLIA Clinical Laboratory Improvement Amendment
Cp Process capability
Cpk Capability index
CRB Center for Biomedical Research
CRM Certified reference materials
Cusum Cumulative Sum
CV Coefficient of variation
LIST OF ABBREVIATIONS (cont.)
DCA Dichloroaniline
DEV% Percentage deviation
DW Distilled water
EQAC External quality assurance in clinical chemistry
EQAC-DLB External quality assessment scheme in chemistry
for district laboratories in Bhutan
EQAS External quality assurance schemes
g Gram
GOD/POD Glucose oxidase /peroxidase
GLDH glutamate dehydrogenase
G6PD Glucose-6-phosphate dehydrogenase
HBsAg Hepatitis B surface antigen
HIV Human immunodeficiency virus
IEQAS International external quality assessment scheme
IFCC International federation of clinical chemistry
IQC Internal quality control
ISO International organization for standardization
IUPAC International union for pure and applied
chemistry
JDWNRH Jigme Dorji Wangchuk National Referral
Hospital
L Liter
LDH Lactate dehydrogenase
LSL Lower specification limit
M Molar
mCpk modified capability index
Mfg Manufacturer
mg Milligram
LIST OF ABBREVIATIONS (cont.)
MIS Misclassification index score
ml Milliliter
MRBIS Mean running bias score
MT Medical technologist
MVIS Mean variance index score
OMVIS Overall mean variance index score
OMRVIS Overall mean running variance index score
ORC Out reached clinic
PAP Phenol and 4-Aminoantipyrine
pNPP p-nitro-phenylphosphate
PPTC Pacific paramedical training centre
PT Proficiency testing
QA Quality assurance
QAP Quality assurance program
QASD Quality assurances and standardization division
RCPA Royal College of Pathologist of Australasia
RE Random error
RIEQAS Randox international EQAS
RM Reference material
SD Standard deviation
SDBIAS Standard deviation of bias
SDI Standard deviation index
SE System error
SOP Standard operative procedures
µl Microliter
TS Total score
LIST OF ABBREVIATIONS (cont.)
U Unit
UK-NEQAS United Kingdom National EQAS
USL Upper specification limit
UV Ultra-violet
VAR Variance
VIS Variance Index Score
WHO World Health Organization
CHAPTER I
INTRODUCTION
1.1 A brief history of the health system in Bhutan
The health practices in Bhutan dates back to the 16
th
century when people
used the knowledge of the various herbs and plants for treating the human sickness but
such practices were very crude and unreliable (1). The modern health care system was
introduced in the 1960s by His Late Majesty Jigme Dorji Wangchuk, the 3
rd
King of
Bhutan. Though the system of modern healthcare is not an old phenomenon, it has
made huge strides into modernity through the decades under the visionary guidance of
the monarchs. With the population of about 0.697 million, an integrated healthcare
delivery system was foreseen as an effective strategy to reach the scattered population
in the rugged mountainous country. The medical system in Bhutan is deeply rooted
and practically linked to the farsighted ground breaking declaration of Gross National
Happiness by His Majesty the King Jigme Singye Wangchuk as the primary objective
of the nation. Lighted by an understanding of the needs of the people to maintain good
health, the government maintains a system of free healthcare for all the citizens of the
country.
In 1961, there were only few modern health facilities in the country but
today, the country has 31 hospitals, 178 Basic Health Units (BHUs) and 485 out
reached clinics (ORCs). Based on the size and location, hospitals are categorized as
national referral, regional referral and district hospitals. The Jigme Dorji Wangchuk
National Referral Hospital (JDWNRH) is the country‟s referral hospital located in the
capital. All the other hospitals are located at various regions of the country with at
least one hospital in each district. At the community level, BHUs and ORCs are
established to provide health services for all the people. Despite the difficulties posed
by various limitations, the health status of the population has been markedly improved
especially during the last 15 years achieving 90% coverage with basic healthcare
services (1, 2).
1.2 Laboratory Services in Bhutan
The medical laboratory services as an important part of health care system
began in 1974 with the establishment of the National Health School at JDWNRH (1).
Later the school was renamed as Royal Institute of Health Science (RIHS) where the
training for the first batch of the general laboratory technicians was began in 1984.
The Central Laboratory at JDWNRH is the centre of medical laboratory services of the
country providing various routine and specialized diagnostic services in different
disciplines of health care services. There are a total of 31 laboratories including one
central laboratory, two regional laboratories and 28 district laboratories. With the rapid
advancement of the healthcare system in Bhutan, laboratories play an important role in
providing the reliable results that are needed for the diagnosis of diseases as well as
monitoring of the treatments.
In order to ensure reliable results and to standardize laboratory methods,
the JDWNRH first began its participation in the Royal College of Pathologist of
Australasia Quality Assurance Programs Pty Limited (RCPA QAP) in 1974.
Furthermore, the Ministry of Health in accordance with policy objectives of the
government, the concepts of Quality Assurance (QA) in health care was outlined and
the program was instituted as the Quality Assurance and Standardization Division
(QASD) in 2001(3). Henceforth, the QASD has been striving for the continuous
improvement in the quality of health care through various activities.
Some overseas laboratory organizations also showed interest for the
improvement of the quality of laboratory services in Bhutan. The survey of pathology
and laboratory services was conducted by a team from Pathologists Overseas based in
California, USA in 2003(4). Survey was conducted to collect information on current
laboratory activities to recommend some steps to improve the quality of laboratory
services. In accordance with the recommendations, clinical laboratory standards were
established in 2008 by the QASD (5). However, like many laboratories in other
developing countries, laboratories in Bhutan face numerous limitations. The inter-
country workshop conducted in 2006 by WHO in Thailand for the South East Asian
Countries reported that, an absence of national standards, a shortage of human
resources, inadequate training facilities and laboratory equipments, and an erratic
supply of reagents as major constraints faced by laboratories in Bhutan (6).
Unavailability and high cost of commercial QC material is another limiting factor in
effort to establish EQAS as the country imports all reagents and QC materials from
abroad. Geographical terrain in Bhutan is also one of the limiting factors causing
difficulties in transportation and communication.
The data analysis and use of the suitable scoring index is an important
phase in the process of EQAS. The scoring systems are used as performance indicators
for the assessment in EQAS (7). Among various scoring indices used in EQAS, the
Capability index (Cpk) has been introduced which provides an indicator of process
bias as well as variations in assessment of the laboratory performance in clinical
chemistry (8). It is also used to compare laboratory performance of an individual
laboratory with peer groups in EQAS (9). Therefore, our aim in this work is to
implement EQAS in clinical chemistry for district laboratories in Bhutan by preparing
a homemade lyophilized human serum QC material and to monitor the laboratory
performance using VIS and mCpk as performance indicators.
CHAPTER II
OBJECTIVES
1. To study the feasibility of implementing EQAS in Clinical Chemistry
for District Laboratories in Bhutan (EQAC-DLB) by collecting information on current
status of laboratory activities through questionnaires.
2. To prepare homemade lyophilized human serum QC material for
implementation of EQAC-DLB.
3. To monitor the quality of laboratory performance in clinical chemistry
by comparing the performance scores in the initial and at the end of study using
suitable scoring index as performance indicators.
4. To use mCpk to evaluate laboratory performance as performance
indicator along with VIS.
CHAPTER III
LITERATURE REVIEW
3.1 Quality Assurance in Clinical Laboratory
A quality is defined as the totality of features and characteristics of a
product or service that satisfies the needs of a customer. In the laboratory medicine,
Quality Control (QC) is designed to detect, reduce, and correct errors in the laboratory
analytical process. Quality Assurance (QA) is concerned with achieving appropriate
levels in matters such as staff training, management, adequacy of the laboratory
environment, safety, storage, identity of samples, record keeping, and calibration of
instruments and the use of technically validated documents. ISO 15189 provides
requirements for competency and the quality that are particular to medical laboratories
(10). Failure in any of these areas might undermine vigorous efforts to achieve the
desired quality of data. Internal Quality Control (IQC) and EQAS procedures are used
in the clinical laboratory to maintain the quality and high standards of analytical
performance.
3.1.1 Internal Quality Control
The IQC involves an in-house procedure for continuous monitoring of
operation and systematic day-to-day checking of the analytical data to decide whether
they are reliable enough to be released for used by the clinicians. The IQC is
particularly used to detect and monitor random or systemic errors by using suitable
QC rules and charts. The IQC techniques can be used with or without using a QC
material. The Correlation of laboratory tests, Delta check, Bull‟s Algorithm and
Average of Normal (AON) are the examples of IQC without using QC material
whereas Six Sigma and QC charts require QC material. The correlation of the
laboratory test involves comparison of the relationship among the laboratory results;
e.g. direct bilirubin should not be greater than total bilirubin. Delta check emphasizes
on the detection of laboratory errors resulted from the mislabeled specimen or dilution
of specimen due to intravenous fluid. It is done by comparing the results of the current
analysis with that of an analysis of a previous day on the specimen from the same
patient. Sheiner LB et al. (11) elucidated the application of Delta check method to
determine the rate of true positive (actual mislabeled) and false-positive (correctly
labeled but taken as mislabeled) specimen. They concluded that the method identified
only half of the truly mislabeled specimens.
Bull‟s algorithm also called “moving average” or simply as “X-B Mean”
was first established in 1974 by Bull (12). It was initially designed to assess the QC of
the RBC counts used by manufacturers to incorporate inside haematology analyzers.
The term “moving average” refers to the process of calculating a mean for a group of
samples that is relative to the mean of the previous group. This means that X-B mean
is continuously updated as patient samples are run. The AON was first introduced in
1965 by Hoffman and Waid. It is based on the assumption of a stable patient
population. It uses median and the 25
th
percentile and 75
th
percentiles as truncation
limit to set up QC limits. For an each day, the patient results falling in QC limits are
averaged and compared with the mean established on the previous day. If the average
falls outside this QC range, the process is considered to be out of control.
In the concept of Six Sigma, the six Standard Deviation (SD) on each side
of the mean should fit within the allowable limits to ensure very low numbers of
outliers (9). QC chart is a data analysis technique to determine if a measurement
process lies out of statistical control. The concept of QC chart was first introduced by
Walter A. Shewhart in 1931 and later in 1950, Levey and Jennings suggested the use
of the same QC chart in the clinical laboratory (13, 14). Today it is the most common
chart used in all the clinical laboratories worldwide. In 1980s, Westgard established
QC rules using the computer simulation model which ensures high probability of error
detection and low probability of false rejection. The Westgard‟s multirule in QC
commonly used are 1
2s
, 1
3s
and R
4s
for detecting random errors; 2
2s
, 4
1s
, 7
T
, 10
x
for
systemic errors. The Westgard‟s multirule is sometimes too strict which gives high
false rejection rate in analytical runs for the laboratories with good precision using
automation. In such situations, the OPSpecs chart is used to reduce false rejection
because it determines the types of QC rules to apply, the number of QC materials and
the runs required (9). Cusum is another QC chart plotted using the cumulative sum of
differences between the target value and the individual laboratory values or
differences between the values and the average (15). It was originally developed by
Page in 1954. It consists of a target mean (which is either reference value or calculated
as the mean of the laboratory data), upper Cusum limit and lower control limit lines
and C+ & C- lines which represent the difference between the mean of the laboratory
data and the target mean. These lines are used to determine if the process is being
shifted indicated by rising trends. A Cusum chart is more efficient than Shewhart QC
charts in detecting small shifts in the mean of a process. However, it is relatively slow
to respond to a large shift and special patterns are hard to analyze (16).
3.1.2 EQAS
EQAS is a fundamental tool for quality evaluation and improvement in
clinical laboratories which involves an evaluation of laboratory by an outside agency
on the performance of a number of laboratories and on the materials supplied. This is
usually organized on a national or international basis. The objective is to improve and
maintain the analytical quality of the laboratory performance compared to peer
groups. Study has been performed on characterization and classification of EQAS
according to objectives such as evaluation of method and participants‟ bias and SD
(17). Today, the EQAS is important not only for the assessment of participants‟
performance, but also to assess the performance of instruments and method. EQAS
plays a crucial role for laboratories seeking accreditation because participating in
EQAS is mandatory for accreditation (18). Many guidelines have been developed for
improving an analytical quality by establishing and managing the EQAS (19). There
are reports on the use of quality specifications in EQAS of the Centre of Biomedical
Research (CRB) in order to design schemes that can assess the reliability of laboratory
performances and meet the changing needs and quality recommendations (20).
Specific proposals have also been made on how to design and execute EQAS by
international working groups, but there seems to be no consensus on the best strategies
and for setting up quality specifications. Therefore, many countries have developed
their own EQAS within and between the countries. There are many international EQA
organizations with which the clinical biochemistry laboratories are registered and
approved by external regulatory bodies. Some of the countries which organize EQAS
are as summarized in the table 3.1.
Table 3.1 Summary of the eight international EQAS organizations
EQA
No of
Assigned values
Scoring
Ownership
Acceptability
Ref. No.
Organizations
participants
indices
Limit
RCPA-QAP
Private
-1000 labs
Consensus of 30%
Z-Scores
21-23
-300
labs using RM
countries
Clinically based
UK NEQAS
Govt
-8000 labs
Consensus values
VIS system,
24-27
-100
Clinical based
Bias/VAR &
countries
ABC system
Randox EQAS
Private
-16,000 labs
Consensus values
SDI, TS
28-30
-86 countries
State of the arts
& DEV%
CDC-EQAS
Govt
-150 labs
Consensus values
Z-Scores
31- 33
-33 countries
State of the arts
CAP-EQAS
Govt
-31,000 labs
CDC RM values
Z-Scores
34, 35
fixed limit
Bio-Rad
Private
-3000 labs
Consensus values
SDI & CVR
36-38
EQAS
-95 countries
Fixed limit
EQAC,
Govt
450 labs
Consensus values
VIS system
39, 40
Thailand
BV
CBI/CMC-
Private
-2000 labs
Consensus values
VIS system
41-43
EQAS
-2 countries
CCV
3.2 QC Material
The goal of every medical laboratory is to obtain reliable and acceptable
results from patients‟ samples. This is achieved by using an appropriate QC material
in IQC and EQAS according to the objective of the schemes. The EQAS organizer is
solely responsible for the appropriate selection of QC materials in compliance with
international standards and scientific recommendation. Characterizing the behavior of
QC material has been proposed using the suitable graphic approach (44). It was useful
for screening or evaluation of candidate QC materials prior to the actual study to
ensure that they mimic certain types of patients‟ specimens and also provide some
information regarding the specificity of analytical methods. Therefore, as general
requirement, QC material should simulate fresh human serum in terms of matrix and
concentration of analytes. It must be homogeneous, stable and non-infective. There
are two main sources of QC materials, animal and human with possible inclusion of
synthetic source.
3.2.1 The matrix effect on QC material
Matrix refers to the substance or base from which the QC material is
prepared by adding required amount of additives such as spiking materials and
preservatives to make desired levels. A matrix effect is defined as the influence of a
property of the sample, independent of the presence of the analytes, on the
measurement and thereby on the values of the measurable quantity. The sample matrix
includes all the components of a material system, except the analytes to be measured.
Two types of matrix effects may be observed, chemical and physical. Chemical matrix
effects are those due to interfering substances and physical matrix effects are due to
differences in physical properties, such as viscosity and physical interaction of low-
molecular-mass substances with proteins between QC material and patients‟ serum.
For this reason, any difference in behavior between QC material and patients‟
specimens should be identified before the study is begun as the matrix effect tends to
cause analytical bias in EQAS. Many proficiency testing programs use QC material
that are derived from modified, often lyophilized supplemented with nonhuman
materials (45, 46). Studied has been done on the matrix effect of QC material
produced by College of American Pathologists (CAP) with objective to determine
feasibility, the usefulness, and potential problems associated with CAP matrix effect
analytical protocol (47). They demonstrated the matrix and calibration biases for
several of the analytes investigated in CAP proficiency testing. Thus, the laboratories
that perform poorly for QC material may perform well in actual patient‟s specimen
and vice versa. These discrepancies in performance could be due to the matrix effect
of the specimen and the study shows that the matrix effect can be avoided by using
fresh human serum which accurately represents the samples used daily in the clinical
laboratory.
3.2.2 Types of QC material
Basically, three types of QC materials can be manufactured or prepared: 1)
frozen serum (modified or unmodified), 2) stabilized liquid serum, and 3) lyophilized
or freeze dried serum.
3.2.2.1 Frozen QC material
There are two types of frozen QC material: modified and unmodified
frozen sera. Modified sera are those collected and processed using excess sera of
patients‟ specimens or plasma from out-dated blood in blood bank. Unmodified
frozen serum or fresh frozen serum is also known as “off-the-Clot Serum” which is
derived from spontaneously clotted whole blood and immediately separated and stored
at - 70
o
C. Preparation of fresh frozen serum and the results of the comparative study
with stabilized liquid and lyophilized serum have been described (48, 49). Minimum
matrix effect and needless of reconstitution steps are the advantages of frozen serum.
3.2.2.2 Stabilized liquid control material
Liquid QC material is prepared either by stabilizing with chemicals or by
sterilization (45). Chemically stabilized liquid control material is cheap and suitable
for use in both IQC and EQA. A serum remains stable for at least 24 days at 2-8
o
C
and 55 weeks if stored at -15 to -20
o
C (50,51). Stabilization of QC material by
sterilization is achieved by filtering the serum using appropriately selected filters.
Serum prepared this way is stable up to 2-3 weeks at 20-25
o
C because it contains no
viable bacteria. The chemically stabilized control material gives rise to commutability
problems due to interference (47).
3.2.2.3 Lyophilized QC material
Lyophilized or freeze dried serum is widely used in QC procedures in
clinical laboratories as QC material. It is stable for 1-2 years when stored at 2-8
o
C and
several years at -20
o
C or bellow. Although the matrix change caused by lyophilization
is much debated by manufacturers and EQA organizers (47) it‟s the most useful
technique for preparing in-house QC material where long term stability is needed. It
also allows assessment of wide range of analytes in EQAS. Elm RJ et al. compared
liquid QC material with two different lyophilized QC materials for 19 analytes (52).
The study showed that variances for 19 analytes in the liquid control materials were
significantly greater than lyophilized QC. They concluded that lyophilized QC though
requires reconstitution, gives more precision than the liquid QC.
3.3 Tests for sufficient homogeneity
QC material distributed in EQAS should be sufficiently homogeneous with
minimum values of vial-to-vial variations. The QC material failing to fulfill the
homogeneity will produce discordant results which may lead to unfair assessment of the
participant performance. According to the International Union for Pure and Applied
Chemistry (©2006 IUPAC), when variations among the sample prepared in bulk is
negligible, it is said to be sufficiently homogeneous (53). The guideline requires a strict
random selection of at least 10 samples analyzed in duplicates. ISO Guide 35 (54) for the
reference material requires the strict random selection of 10-30 samples measured in three
replicates. IFCC Guidelines on EQAS developed by Hill P et al. (19) recommends the
selection of the first, the middle, and the last vial in the filing line and two vials from the
middle and two from the outer position of the lyophilizer chamber. Each vial needs to
analyze 3 times within the single series of analysis to ensure that all the vials are equally
affected by the possible drift with the time in analytical system. The working group
recommended for the examination for the variation of sodium, calcium, total protein,
bilirubin, glucose and ALT using the one-way ANOVA. If the initial tests show
significant between-vial variation, it can be confirmed by further investigation by
selecting 2% of the total vials covering the critical analytes. The (©2006 IUPAC protocol
requires the sampling variation (S
sam
) to be less than 30 % of the target SD (σ
p
) and
analytical precision should satisfy σ
an
/σ
p
< 0.5 where, σ
an
is the repeatability SD in homogeneity testing. At 95% confidence
interval in one sided tests for sampling variance, the test should satisfy S
2
sam
>F
1
σ
2
all
+ F
2
σ
2
an
, where, F
1
and F
2
are constant derived from standard statistical tables.
Analytical variance or repeatability (σ
2
an
) shows within-vial variation and the
sampling variance (S
2
sam
) shows between-vial variation. The duplicate results have to
be tested for significant difference by Cochran‟s test at 99% level of confidence
interval. Data sets containing discrepancies in two distribution units should be
discarded. The pair of results with outlying mean value but without evidence of
extreme variance should not be discarded.
3.4 Tests for sufficient stability
Materials distributed in EQAS must be sufficiently stable over the period
in which the assigned value is valid. The term “sufficiently stable” implies that there is
no significant variation during the entire period of study and without adverse effect on
the interpretation of the results (53). According to ISO Guide 35 for the reference
materials (54), two conditions are required to select for stability study, storage
conditions for long-term stability and transport conditions for short-term stability.
Short term stability study usually takes 1-2 months and requires 3-5 data points over
two weeks from 6-10 vials. The long term studies which lasts 24-36 months requires
5-6 data points in time from at least 10-12 vials for each temperature. The study on
stability should be preferably conducted under repeatability conditions to obviate the
effect of reproducibility effect on uncertainty estimation. According to the ©2006
IUPAC protocol for PT, stability test involves the comparison of the analyte between
materials subjected to extreme conditions and normal conditions which are labeled as
“experimental” and “control”, respectively. In this concept, distributed samples are
randomly divided into equal subsets. The experimental subset is subjected to the
appropriate treatment (e.g. high temperature), while the control subset is kept under
conditions of maximum stability (e.g. low temperature). The materials are then
analyzed simultaneously for mean difference using independent Student t-tests with
pooled standard deviation at 95% confidence interval. Any highly significant
difference between the mean results of the two subsets can then safely be regarded as
evidence of instability.
3.5 Evaluation of EQAS Results
Data analysis is an important phase in the process of EQAS. The data
reported by the participants are analyzed by providers using a suitable statistics which
is the main tool for the assessment of the laboratory performance. According to the
WHO EQAS Guideline for health laboratories (7), the main objectives for analysis of
EQA results are i) to provide an overall summary of the laboratory performance; ii) to
assess each laboratory, the current and previous performance in the EQA survey; iii)
interlaboratory comparison of performance with adequate number of participants
using the similar methods and instruments. The minimum data to be reported to the
participants are the number of laboratories in the method group, the overall and
method group mean and SD, the results submitted by the participants, assigned values,
acceptable limits and the performance scores.
3.6 Assigned values for EQAS
There are many approaches for the EQAS provider to determine the
assigned value and its uncertainty (53-56). Selection of the method to establish
assigned values differs from scheme to scheme or even round to round within a
scheme depending on the purposes and design of the schemes. However, there are
many guidelines published by the committees of the well-known EQA providers,
reference laboratories and accreditation bodies that set certain requirements to fulfill in
establishing assign values. The statistics to be applied in establishing assign values and
acceptability criteria is covered in ISO 13528-2005 (55) and ISO/IEC Guide 43-1E
(56). The approaches commonly applied are: consensus of laboratories accredited to
ISO 15189; measurement by a reference laboratory; use of a certified reference
material (CRM); direct comparison of the EQAS results with values of CRM;
formulation approach and consensus of participants. Derivation of assign values using
consensus of laboratories accredited to ISO 15189 involves distribution of samples to
minimum of 3 laboratories. The assigned value is taken as the consensus of those
laboratories. This method is useful when operationally defined parameters are
measured by participants whose results are expected to be consistent with results from
a smaller number of laboratories. It provides forum for cross-checking among the
laboratories, thereby prevent gross errors.
3.7 Limit of acceptability
Limit of acceptability is the important benchmark for the interpretation of
the laboratory results. It provides the scale for the laboratory deviation and used to
derive criteria for performance evaluation usually expressed as percentage of the
assigned values. A performance standard established in this manner after taking into
account whether reported results are fit for the purpose is known as “analytical goal”.
There are many approaches to establish the analytical goals for EQAS (57-59). Ricos
et al. have categorized the limits of acceptability used in European countries as “fixed
limit” and “variable limit” (57). Rhoads (58) has defined the analytical goals based on
clinical application, establishment and deployment. Clinically, the analytical goal is
defined as “the amount of error that can be tolerated without invalidating the medical
usefulness of the analytical result. Deployment definition defines a set of rules which
convert a total allowable error into target SD for daily QC and allowable bias.
Generally, analytical goals are defined based on the approach made in establishing the
analytical goals: medical requirements; biological variation; reference interval;
regulatory requirements (CLIA ‟88) and consensus of participants.
3.8 The performance scores
The scoring systems are used for the judgment of individual laboratory
indicators for the assessment of the overall performance in EQAS (7). Scoring indices
provide common method for data quantification that is applicable to all the tests which
are least or not influenced by methodological conditions, reagents and instruments
(60). The scoring system depends on the types of tests offered by the scheme
providers, i.e. qualitative, semi-quantitative and quantitative. Scores also differ from
scheme to scheme depending on the scheme design, quality goal and specification. For
the qualitative and semi-quantitative tests, misclassification index scores, credit scores
and penalty scores are used for performance evaluation. For the quantitative tests,
SDI, Z-score, VIS, Bias index score (BIS), Bias & VAR scoring system, ABC system
and Cpk has been recently introduced to be used in interlaboratory comparison (10).
VIS was first proposed for the UK-NEQAS (26). The deviation of a laboratory's result
in respect to a target is expressed as a standard score. It is defined as the difference
between the result obtained by laboratory and the calculated method mean expressed
as a percentage of the mean divided by CCV for that determination. The resultant
figure is multiplied by 100 to avoid decimals and the sign is ignored. VIS is also
defined by the formula: VIS = {[(X
lab
-AV)/AV x100]/ (CCV) x100} where, X
lab
is
the individual analytical values for each analyte, AV is the assigned value after
removing the outliers. CCV is the chosen coefficient of variation which is a scaling
factor for each analyte correcting for the differences in the state of the art yielding VIS
in common currency. UK-NEQAS and WHO-EQAS VIS as: VIS < 50 -excellent; 50
< VIS < 100-very good; 100 < VIS ≤ 150-good; 150 < VIS ≤ 200-satisfactory and
VIS>200-unacceptable. The maximum VIS is kept as 400 to avoid inclusion of very
high VIS resulted from the random and clerical errors. The mean VIS for several
distributions of material using all the results of the same determination can be
calculated to compare with the mean of VIS by all the participating laboratories called
MVIS which is interpreted similar with VIS. The mean VIS calculated by including
different types of determinations called Overall Mean VIS (OMVIS). OMVIS can be
used to assess the overall laboratory performance and on individual determinations.
The mean VIS for the most recent 40 results returned by the individual laboratory
which includes variety of determinations is called Overall Mean Running VIS
(OMRVIS). It is the most useful performance indicator to assess changes in the
performance of individual laboratory with time. The mean VIS of the 10 most recent
VIS can be calculated to obtain MRVIS. MRVIS is the indicator of total errors and
allows the identification of the analytes with greatest difficulties relative to the state of
the arts (35). Thus VIS scoring system is a simple and robust to indicate both bias and
variation in the test system which may be used as a practically useful indicator of the
laboratory‟s general standard of performance and of changes with time it has been
extremely useful in many national and international schemes. The concept of process
capability has been used to quantify the relationship between product specifications
and the measured process performance (8). Of the various available capability indices,
the Process Capability (Cp) had been widely used. Cp is defined by the formula: Cp =
(USL-LSL)/6SD where, USL and LSL are the upper and lower specification limit of
an analytical process, respectively which corresponds to the formula, TEa/6SD, where
TE
a
is total allowable error and 6SD (±3SD). Burnett L et al. performed an experiment
to assess whether Cp alone can be used to select appropriate QC rules. Their results
demonstrated that the performance of the multichannel clinical chemistry analyzer
improves with the use of different QC rules selected on the basis of Cp. However, this
approach has some disadvantages. Cp is the most basic of family capability indices
which includes consideration of only imprecision, but not of bias. Other indices such
as Cpk include consideration of bias as well as imprecision. Cpk is defined as Cpk =
(TE
a
– Bias)/3SD where, TE
a
is the total allowable limits of error, SD the standard
deviation, and bias is the average deviation of the laboratory method in comparison to
assigned value or reference mean. Therefore, Cpk provides an indicator of process bias
as well as precision. The acceptable limit of the scores is interpreted as follows:
Cpk < 1: Incapable, 1< Cpk ≤ 3: Capable and Cpk > 3: Very capable. Bais R
determined the Cpk from Royal College of Australasia Quality Assurance Program
(RCPA QAP) end-of-cycle summary data using the allowable errors from the RCPA
QAP (9). The statistical result obtained by using the formula, Cpk = TE
a
– Bias)/SD,
demonstrated that Cpk can be used to detect errors where the SD is widely compared
to clinical requirements. In EQAS, Cpk can be used for monitoring performance on
individual analytes, comparing laboratory performance with peer groups and to
optimize the amount of quality control run in a day.
CHAPTER IV
MATERIALS AND METHODS
4.1 General materials
4.1.1 Equipments
The following equipments were used from the Faculty of Medical
Technology, Mahidol University:
Lyophilizer, Dura-stop/MP, Cat No. FD20A001, FTS System, Inc., USA
Chemistry analyzer (Hitachi 917) Cat. No. 0727-20, Roche Diagnostic, USA.
Refrigerator: Sharp, Thailand.
Deep freeze (-20
o
C): Model FC-27, Sharp, Thailand.
Automatic Dispenser: oxford laboratories, USA.
Centrifuge: Cat No. 8406543, SORVALL, USA.
PH meter: Cat No. 9608, Beckman, USA
Weighing scale: Cat LP-2245, Scientific Promotion Co. Ltd., USA.
Vacuum pump: Cat No. DOA-V153-BN, Benton Harbour MICH,
USA. Adjustable micropipette, Model M-5000DG, Nichiryo Co. Ltd.
Japan. Magnetic stirrer plate: Model MT2, Amicon, USA.
Filer, 20µm pore size, AP20 Cat. No. 07500 Dia 75mm, Millipore Co. USA.
Filter membrane 0.45 µm, AW06, Cat. No. 09000 Dia 90mm, Millipore Co.
USA.
Stainless steel filter holder, milipore, USA.
Brown glass vial 3 ml, 7x13mm mouth, Cat No. 223684, BIOMED Co.
Germany.
Gray Stopper, 7x13mm, Cat No. 224100-081, BIOMED Co. Germany.
Computer: AMD Turion 64x2 Mobile Technology, TL-58 processor, 3
GB Ram, 160 GB harddisk, Softech Company, Thailand.
Centrifuge tubes (250 ml), Cat. No. 340198, Beckman, USA.
The general glassware like, beakers, measuring cylinder, volumetric
pipettes and Test tubes used in this study were from Pyrex
®
, USA, Kartall,
Milano, Italy and witeg in Germany.
4.1.2 Chemicals & Reagents
Glucose, Cat. No. 112H0142, Sigma, USA
Urea Cat. No. 73H0952, Sigma, USA
Creatinine, Cat. No. 63H0920, Sigma, USA
ALT, Cat. No. 107K7002, Sigma, USA
ALP, Cat. No. 039K1213, Sigma, USA
Total bilirubin, Cat. No. 055K0919, Sigma, USA
Hydrochloric acid, Cat No. K40132917-918 MERCK, Germany
Sodium hydroxide, Sigma, USA
4.1.3 Miscellaneous items
Plastic bottles, 250 ml
Whisk & funnel
Carrier basket
Blue tips
Watmann filter paper No. 1
Mailing envelops
Labeling stickers
Packing materials (14x20 cm box)
4.2 Methods
4.2.1 Study on the feasibility of implementing EQAS.
4.2.1.1 Collection of information on health systems using
the official websites.
There are total of 31 hospitals in Bhutan. Based on the bed
size and the location, all the hospitals are included in our study to find out the number
of laboratories performing the clinical chemistry tests (figure 4.1).
Figure 4.1 Map of Bhutan showing the location of hospitals and laboratories in
Bhutan. NRH-National Referral Hospital; RRH-Regional Referral Hospital; DH-
District hospital.
4.2.1.2 Collection of information using questionnaire
Questionnaire was prepared for the entire hospital laboratories
t collect information on current status of laboratory activities in Bhutan. Our
questionnaire was divided into five sections based on the type of information required
for the study. Section 1: the general information: the category, number of beds and
location of the hospitals. Section 2: Laboratory information: reagents, methods,
instruments and types of tests performed. Section 3: Educational information: training
for laboratory quality assurance, number of staff and qualification. Section 4:
Information on QA: participation in EQAS, types of IQC materials, levels and design
of IQC, handling and storage of the QC materials, equipments used for reconstitution
and Levey-Jennings chart to monitor the IQC. Section 5: Basic knowledge on safety
and precautions in the laboratory. The questionnaire was sent to the coordinator at the
Department of Central Laboratory in Thimphu via e-mail. Coordinator was requested
to fax the questionnaire to the laboratories immediately and informed to return it back
within a week. After completing the questionnaires, all the laboratories returned it to
the coordinator by fax and the coordinator sent to the investigation team via speed
post and email. The completed questionnaires returned by the participants were
analyzed the results were used for the next plan of activities.
4.2.1.3 Assessment of basic QA knowledge in laboratory
personnel in Bhutan
We planned three days workshop to impart participants with
basic knowledge on laboratory QA before implementation of our project. The
workshop was organized by the coordinator, where the heads of the laboratories were
invited. The concept of EQAS and IQC in clinical chemistry was the theme of the
workshop. They were briefed on reconstitution of control material, calibration of the
photometers and use of Levey-Jennings chart. Pre-tests and post-tests were conducted
to determine the effectiveness of the workshop on QA. The test papers were evaluated
and the performance scores were given to each participant based on their performance.
4.2.2 Preparation of control material
4.2.2.1 Collection of specimen
Following an informed consent, about 300 ml of fresh whole
blood was collected from each Bhutanese volunteer using a standard procedure at the
Faculty of Medical Technology. Serum was separated and about 1 ml of serum from
each donor was sent to the Centre of Medical laboratory Services, Faculty of Medical
Technology, Mahidol University to check for anti-HIV and HBsAg. The test results
were negative for anti-HIV and HBsAg. The separated serum was kept frozen at -
20
o
C until it was processed on the next day for lyophilization.
4.2.2.2 Specimen processing
In this study, we followed the protocol described in the WHO
document/LAB/81.4 to prepare lyophilized human serum control material. After the
negative test was declared for anti-HIV and HBsAg, frozen serum was thawed at room
temperature for 1 hour and pooled in a flask. The pooled serum was then carefully
mixed to ensure homogeneity using the magnetic stirrer, then centrifuged at 2000g for
10 minutes and filtered with Watmann filter paper No. 1 to minimize the possibility of
adding contaminants in the process.
4.2.2.3 Level adjustments
In order to prepare the serum concentration of the desired
levels, pooled serum was divided into four parts: 2 parts for normal (N) and the other 2
parts for abnormal (A). To achieve complete blinding of the participants, both normal
and abnormal levels were divided into 2 sub levels designated as N
1
and N
2
for
normal and A
1
and A
2
for abnormal, respectively. The baseline values obtained for
each analyte was kept as the N
1
. 35 ml of distilled water was added to N
1
to obtain N
2
using the formula, V
1
C
1
= V
2
C
2,
where V
1
& V
2
are initial and final volume and C
1
& C
2
are baseline and expected values, respectively. To obtained abnormal levels,
carefully calculated amount of selected biological compounds were measured and
added as shown in the table 2 to prepare A
1
and A
2
.
Table 4.1 Level adjustments by adding selected biological compounds.
Normal level (N): 35 ml
Abnormal (A) : Required amount added
D/W added to N
1
to get
Level: A
1
Level: A
2
N
2
Tests
Unit
Value
Value
Baseline
TV
Obtaine
TV
Amount
Value
TV
Amount
Obtaine
level(N
1
)
*
d
*
added
obtained
*
Added
d
(N
2
)
Glucose
mg/dl
80
75
74.0
120
0.14
123
150
0.022
145
BUN
mg/dl
13
12
11
30
.053
20
40
0.08
23
Creatinine
mg/dl
0.9
0.8
0.8
2.0
0.0034
2.2
2.5
0.005
2.7
AST
U/L
28
25
24
-
-
21
-
-
19
ALT
U/L
25
22
22
90
0.32
94
200
0.77
222
ALP
U/L
99
88
88
290
1.60
347
320
1.80
398
T. Bilirubin
mg/dl
0.84
0.74
0.71
2.0
0.0051
2
2.5
.0053
2.5
T. Protein
g/dl
7.9
7.0
7.0
-
-
7.8
-
-
7.7
Albumin
mg/dl
4.9
4.3
4.4
-
-
4.9
-
-
4.9
Dispensing
Total of 560 vials were divided into four groups for 4 levels of
samples, N
1
, N
2
, A
1
and A
2.
Samples from each flask were dispensed into 3ml amber
colored vials with gray stopper using the automatic dispenser. Vials were grouped into
10 vials and tied around with colored rubber bands to identify the four levels and
properly arranged on the lyophilizer tray.
Lyophilization
The tray containing the sample vials were transferred to the
lyophilizer chamber and a temperature probe was placed in one of the vials to monitor
the drying process and this vial was at the core of the batch. The serum was
lyophilized for 18 hours until the monitor displayed the sign as “END” on the next
day. As soon as the lyophilization was completed, the tray was removed out from the
lyophilizer and vials were capped immediately, sealed with plastic ring and labeled
with explicit information. The overall work flow for the preparation of lyophilized
sample is shown in figure 4.2.
Figure 4.2 Work flow for preparing homemade lyophilized human serum for
implementing EQAC-DLB.
4.2.3 Study on the characteristics of the control material
4.2.3.1 Homogeneity testing
To test for the vial-to-vial variations, samples were tested for
homogeneity by following the ©2006 IUPAC Protocol. 7 vials were randomly
selected. The first, the middle, and the last vial in the filling line were selected as well
as two vials from the middle and two from outer positions of the lyophilization
chamber. Each sample was measured in duplicates. The results were first checked for
outliers using Cochran‟s t-test and compared with critical values at 95% and 99%
confidence interval which were 0.727 and 0.838, respectively (53). No outliers were
detected and the results were further tested for significant sampling variance.
Sampling variances were calculated and compared with critical F value at 95%
confidence interval.
4.2.3.2 Establishment of assigned values
Three samples of each level were analyzed in duplicate at
three ISO 15189 certified laboratories in Thailand for three days. The results from
each of the three laboratories were tested for outliers and robust mean and SDs were
calculated as described in ISO 13528. The chemistry analyzers used by the 3
laboratories were Hitachi 917 from Japan, Modular P800 and Cobas Integra
®
400 from
Roche Diagnostic Co. Thailand. The degree of variation of all participants‟ results
from the assign mean were assessed by using VIS scores calculated using CCV from
CLIA PT criteria as the scaling factor.
4.2.3.3 Stability of control material
Total of 40 samples were randomly selected and they were
divided into 3 groups to store at 3 storage conditions (i.e. at room temperature, 2-8
o
C
and at -20
o
C) to study the stability at these temperatures. 5 vials each of N
1
and A
1
are
stored at room temperature and another 5 vials each of N
2
and A
2
are stored at 2-8
o
C.
5 vials each from all the four levels are stored at -20
o
C. They were analyzed in 3rd,
5th, 7th, 8th and 9th month after the date of preparation. Each of the samples was
analyzed in duplicate. At the end of five rounds, the results were compared with the
initial values generated in the homogeneity tests. We used SPSS program to test for
equality of variance by Levene‟s test and t-test for equality of the means with 8
degrees of freedom at 95% confidence interval.
4.2.4 Implementation of EQAS in Bhutan
4.2.4.1 Allocation of the samples
There were initially 19 laboratories to analyze two samples at
a time in each month for 12 rounds. Therefore, lyophilized serum was allocated for the
following purposes: 516 vials for analysis by district laboratories in Bhutan, 36 vials
for three ISO certified laboratories in Thailand to determine the target values, 28 vials
for homogeneity tests and 40 samples for stability tests.
4.2.4.2 Result recording forms and package inserts
Data recording form and package insert with instruction on
reconstitution technique that accompany the specimen was prepared in a simple way to
understand (appendices B & D). For identification of results from different
laboratories, each vial were labeled carrying the information of the work, date of
preparation, lot number, sample code, sample serial number and the recommended
storage conditions and volume of reconstitution fluid to be added. Instructions in the
package insert were: handling the control material; procedure for reconstitution;
storage of control material; result recording and Safety & precautions.
4.2.4.3 Design of the scheme for monthly analysis of QC
sample
The trial number starts from 001 to 024 for each hospital (table
4.2). Two trials were scheduled to be analyzed in each month the first being trial 001
and 002. The sample levels were randomly distributed in 12 months in order to ensure
complete blinding of the participants i.e. measurement of same levels of control
sample in a month was avoided by providing a pair of samples containing two
different levels in each month. The 24 trial numbers were randomly assigned for 4
level groups of the sample as follows: i) N
1
: 001, 006, 009, 013, 018, 021; ii) N
2
: 003,
008, 012, 015, 020, 024; iii) A
1
; iv) 002, 005, 011, 014, 017, 023; iv) A
2
: 004, 007,
010, 016, 019 and 022.
Table 4.2 Design of the EQAS rounds with trial numbers
Month
Nov
Dec
Jan
Feb
Mar
Apr
May
Jun
Jul
Aug
Sept
Oct
Year
09
09
10
10
10
10
10
10
10
10
10
2010
Trial No.
001
003
005
007
009
011
013
015
017
019
021
023
Levels
N1
N
2
A1
A
2
N
1
A
1
N1
N2
A1
A
2
N1
A1
Trial No.
002
004
006
008
010
012
014
016
018
020
022
024
Levels
A1
A
2
N1
N
2
A
2
N
2
A1
A2
N1
N
2
A2
N2
4.2.4.4 Distribution of the QC sample to the participants
In view of anticipation for transportation problem in Bhutan,
we prepared samples for single mailing for the entire period of study. Paired samples
were wrapped in plastic sheets and packed in the protective container containing 12
pairs of sample for 1 year for each hospital. They were properly labeled with safety
information, storage conditions and address of the hospitals. Total of 20 packages
were packed together in the thick protective box to ensure safety in transportation. One
package containing extra control samples were sent to replace in case of problem.
They were placed in refrigerator at 2-8
o
C till it was dispatched to the coordinator by
flight to Jigme Dorji Wangchuk National Referral Hospital in Thimbu. The
coordinator was asked to keep the samples inside freezer or refrigerator until
dispatched to the participants. The means of sample transport from Thailand to Bhutan
is as shown in figure 4.3.
Figure 4.3 Flow diagrams for the sample distribution from Thailand to Bhutan and to
the participants of EQAC-DLB.
4.2.4.5. Performance monitoring and feedback reports
Every month the two control samples were analyzed by
participants in the first week of the month. Participants were asked to send the results
within a week after analysis. Their results were sent to the coordinators at the central
laboratory in Bhutan by fax and then to the investigator via email. Total of 12 rounds
of analysis were initially planned (table 4.2). However, we shortened the period by 2
months on logistic grounds. Their results were statistically analyzed and feedbacks
were sent immediately with suggestions for improvement for those with extreme
outliers through the same route of communication. Our monthly feedback report
consisted of overall mean, SD and VIS scores for each analyte (appendix C).
Distributions of the results were depicted using the histogram to compare the
performance among individual laboratory on each analyte.
Table 4.3 Schedule for sample analysis and reporting for EQAC-DLB.
Sample analysis and results
Result analysis by
Dispatch of the
overall
reporting by participants
investigators
Feedback reports
Assessment
20 -25
th
Nov, 2009
26
th
Nov -10
th
Dec, 09.
15
th
Dec, 2009
1-10
th
Dec, 2009
11
th
–20
th
Dec, 2010
25-30
th
Dec, 2010
1-10
th
Jan, 2010
11
th
–20
th
Jan,2010
25-30
th
Jan, 2010
1-10
th
Feb, 2010
11
th
–20
th
Feb, 2010
25-30
th
Feb., 2010
1-10
th
Mar, 2010
11
th
–20
th
Mar, 2010
25-30
th
Mar, 2010
1-10
th
April, 2010
11
th
–20
th
April, 2010
25-30
th
Apr, 2010
5-15
th
1-10
th
May, 2010
11
th
– 20
th
May, 2010
25-30
th
May, 2010
August, 2010
1-10
th
June, 2010
11
th
– 20
th
June, 2010
25-30
th
Jun, 2010
1-10
th
July, 2010
11
th
– 20
th
July, 2010
25-30
th
July, 2010
20-25
th
July, 2010
26
th
–30th
th
July, 2010
5
th
Aug, 2010
1-10
th
Sept, 2010
--Project ended--
--Project ended--
1-10
th
Oct, 2010
--Project ended--
--Project ended--
4.2.5 Performance assessment
The results reported by the participants were statistically analyzed. Overall
data analysis was performed to assess the laboratory performance by district
laboratories in Bhutan. The following assessments were included in this study as
follows:
4.2.5.1 Assessment of interlaboratory precision by
comparing overall monthly CV with CCV from CLIA PT criteria
The overall mean, Standard deviation (SD) and coefficient of
variation (CV) were calculated for both the sample trials after removing the outliers
using ±2 SD. CV for each analyte is compared with CV from the CLIA proficiency
testing criteria which is also used in this study as the chosen coefficient of variation to
calculate variance index score (VIS).
4.2.5.2 Evaluation of performance using VIS system
VIS for each analyte of the results from the participants was
calculated by applying the formula: VIS = [(X
lab
-AV)/AV x 100]/ CCVx100, Where,
X
lab
is the result from the participants, AV the Assigned values from ISO 15189
certified laboratories and CCV is the chosen coefficient of variation from CLIA PT
criteria. For our participants we interpreted VIS as follows:
VIS: 0-50 - excellent performance
VIS: 51-100 - very good performance
VIS: 101-150 - good performance VIS:
151-200 - acceptable performance
VIS: 200-250 - performance needs improvement
VIS > 250 - unacceptable performance, needs immediate
action. The acceptable VIS Scores for our participants has been kept higher by 50
scores than required by UKNEQAS considering the manual methods used by majority
of our participants. The maximum VIS is up to 400 and the sign is ignored. The mean
of the VISs for entire studies for each analyte and level of the material was calculated
and is called MVIS. Taking the mean of all the VIS irrespective of the types of
analytes and levels gives the overall mean VIS (OMVIS) which is a useful score to
assess overall performance. Following assessments were performed using the MVIS:
1. Assessment on instrument and method performance using MVIS
2. Comparison of performance on four levels of control material
3. The overall performance evaluation by grouping VIS into three
categories A, B and C. Interpretation: A= VIS < 200 as acceptable; B = VIS: 200-250
Needs improvement; C = VIS>250 as unacceptable.
4.2.5.3 Evaluation of performance using index mCpk
We used the Cpk defined by the formula Cpk = (TE
a
–
Bias)/3SD Where, TE
a
is the total allowable limits of error; SD, the standard deviation
of between-batch QC measurements and Bias is the average deviation of the overall
mean from the assigned value. However, we found some disadvantages in using this
scoring index in evaluating the results when overall data contains high SD and bias
which is expected from the laboratories using manual methods in developing
countries. This high imprecision generates very small Cpk values with one or two
decimal points causing inconvenience in evaluation and reading of the scores. We
modified a Cpk using the concept of total allowable error specified by the CLIA PT
criteria for the analytical requirements and capability index proposed by Renze Bais
(9) as follows:
TEPT = Biasmeas+ Biasmatx + SEcont Smeas+ zREcont Smeas
(1)
Simplifying the formula (1),
TEpt-(Biasmeas+ Biasmatx ) = Smeas (SEcont + zREconts )
(2)
Where, Bias
meas
is the systematic difference between results
by the measurement procedure and the true values; Bias
matx
accounts for any
systematic difference for the specimen matrix due to interference or lack of specificity
of the measurement procedure; S
meas
is the SD of the measurement procedure; SE
cont
is the systematic error and RE
cont
is the random error that can be detected by the QC
procedure; z is a multiplier related to the portion of the distribution exceeding the
quality requirement which is set as 1.65 to fix the maximum defect rate as 95%
confidence interval assuming a normal Gaussian distribution. We designated, S
meas
(SE
cont
+ zRE
cont
) as the combined errors (CE) of the measurement and simplified the
formula:
Combined errors (CE)=TE
a
- Bias
(3)
CE is expressed in % of the mean
%CE = (TE
a
– Bias)/X
lab
x 100
(4)
Then we divided % CE by laboratory CV to obtain mCpk and
multiplied by 100 to make mCpk in common currency and here the signs are retained
which reflects the magnitude of the bias.
mCpk = [(TEa–Bias)/X
lab
x100]/CVx100 (5)
The performance is graded using Cpk as follows:
mCpk < 200 –Incapable performance
mCpk: 200 – 300 – Fairly Capable
mCpk: 301 – 400 –Capable
mCpk > 400 -Highly Capable
The scores are interpreted as summarized in table 4.4.
Table 4.4 Interpretation of the modified capability index (mCpk)
Bias
CV
mCpk
Interpretation
Acceptable
Acceptable
High positive
capable
Unacceptable
Acceptable
High negative
Inaccurate
Acceptable
Unacceptable
Low positive
Imprecision
Unacceptable
Unacceptable
Low negative
Incapable
Bias = TEa
Any
Zero
Inaccurate
4.2.5.4 Assessment of the EQA in relation to the IQC of the
participants
The results of the IQC performed along with our sample were
collected from 7 selected laboratories. The mean, SD and the range at ± 2SD were
calculated. The CV was calculated to determine the precision of the laboratories. The
glucose and AST results of IQC and EQAS from seven laboratories for 6 selected
trials were used to assess the laboratory performance in relation with performance on
IQC. The IQC results were compared with corresponding EQA results of the same
day. The EQA results of the sample level A
1
and N
1
from seven laboratories were
used as inferential data to evaluate the laboratory performance on IQC and EQA
results. The data shown here was examined for the CV greater than CCV according to
the WHO guidelines. EQA results from each laboratory were compared with the range
of the assigned value at mean ±3SD. The overall work plan for the study is as shown
in the figure 4.4.
Figure 4.4 Overall work plan for the activities carried out to implement EQAC-DLB.
CHAPTER V
RESULTS
5.1 Results from the study on the feasibility of implementing EQAS
5.1.1. Information on health systems using the official websites
There are 31 hospitals, 178 Basic Health Units (BHUs) and 485 out
reached clinics (table 5.1). BHU-I is provided with one medical doctor along with
about two nurses and one clinical laboratory assistants. BHU-II is managed by
assistant clinical officer accompanied by an assistant nurse. Based on bed size and
location, hospitals are categorized as central, regional and district hospitals. All the
district laboratories and Basic Health Units (BHU) up to level I perform basic
laboratory tests in haematology, urine microscopy, rapids tests in immunology.
Clinical chemistry tests are performed by only some of the district laboratories but not
at all the levels of BHUs.
Table 5.1 Current Health Centres in Bhutan
Type of hospital
Bed size
No of centres
National Referral Hospital
350
1
Regional Referral Hospital
150
2
District Hospitals
20-40
28
Basic Health Unit 1 (BHU-I)
5-10
15
Basic Health Unit II (BHU-II)
3-5
163
Out reached clinics (ORCs)
0
485
5.1.2 Information from the questionnaires
The information collected regarding the category of hospital with bed size,
number of laboratories using various equipments is summarized has been summarized
in table 5.2. Hitachi 911 and Beckman Coulter CX5 are the automated analyzers used
by central and one of the two regional laboratories. Smart lab is used for glucose
measurement by one of the regional laboratories. The district laboratories use
semiautomatic photometers among which eight laboratories use BA-88 and nine
laboratories use Screenmaster-3000.
Table 5.2 The bed size of hospital and number of laboratories in Bhutan using
various clinical chemistry analyzers
Category of hospital
Bed Size
Analyzers for chemistry tests
No of
laboratories
Central laboratory
350
Hitachi 911
1
Regional laboratories
150
Beckman Coulter CX5 & Smartlab
2
District laboratories
20-40
Photometers
16
Total laboratories performing clinical chemistry tests in Bhutan
19
District laboratories perform only routine basic tests including glucose,
renal function test, liver function test and proteins. The special tests such as cardiac
marker, tumor markers, hormones, etc. are provided only at central laboratory and one
of the regional laboratories. The regents and control materials are commercially
purchased. Central laboratory uses reagents from Diasys and Trulab N and Trulab P as
IQC (table 5.3). One of the Regional Referral Hospitals uses the same reagent for
Smart Lab and Beckman Coulter reagent for CX5 PRO with Synchron LX
®
DxC as
IQC. District laboratories use reagents from Coral Clinical Company and precinorm
and precipath as the IQC. The central laboratory and one of the regional laboratories
use different reagents and methods whereas all the district laboratories use same
reagents.
Table 5.3 Methods/principle and volume of sample required per run for the
laboratories in Bhutan.
Automated analyzer
Semi automated
Tests
Hitachi 911/Smart Lab
Beckman Coulter CX5
Photometers
Methods
S. Vol
Methods
S. Vol
Methods
S. Vol
Glu
GOD/PAP
6 µl
HK-G6PD
10 µl
GOD/POD
10 µl
BUN
Urease GLDH
5 µl
UV kinetic
303
Bertholet
10 µl
µl
Creat
Jaffe
15 µl
Rosaliki method
23 µl
Jaffe
100
µl
AST
UV kinetic
10 µl
UV kinetic
23 µl
IFCC method
100
(IFCC)
(Henry)
µl
ALT
UV kinetic(IFCC)
15 µl
UV kinetic(Henry)
23 µl
IFCC method
100
µl
ALP
pNPP Kinetic
4 µl
pNPP Kinetic
5 µl
pNPP Kinetic
10 µl
T.B
DCA Method
7µl
Jendrassik Grof
8 µl
Jendrassik
100
Grof
µl
T.P
Biuret Method
6 µl
Biuret Method
300
Biuret method)
20 µl
µl
Alb
BCG
6 µl
BCG
300
BCG
10 µl
µl
Table 5.4 The manufacturer’s range used as IQC range by laboratories in
Bhutan
Central laboratory
Regional laboratories
District laboratories
Tests
Units
Trulab N
Trulab P
Low
Normal
High
Precinor
Precipath
m
Glu
mg/dl
74.3-103
241-333
39-55
215-255
386-454
80.0-
222-300
108.2
BUN
mg/dl
30.0-46.8
116-182
13-21
68-86
125-151
0.90-1.26
3.17-4.55
Creat
u/l
0.70-1.10
6.27-9.81
0.3-0.7
3.7-4.5
6.9-8.5
38.6-55.4
116-164
AST
u/l
14.4-23.0
65.9-105
24-34
147-198
276-356
38.9-56.3
101-149
ALT
u/l
55.4-92.3
175-291
24-34
138-198
257-357
66.3-95.7
191-275
ALP
mg/dl
0.84-1.64
4.99-8.49
30-50
121-171
211-291
0.91-1.33
3.92-5.66
T.B
g/dl
4.40-5.48
5.85-7.29
0.5-1.5
3.1-5.3
5.4-9.4
6.07-7.75
4.28-5.48
T.P
g/dl
2.88-4.60
3.54-5.66
3.2-4.2
5.3-6.5
7.3-8.9
3.98-5.72
2.53-3.61
5.1.3 Laboratory staff and qualification
Only the central and regional laboratories have the pathologists and
medical technologists (table 5.5). Rests of the laboratories are operated by clinical
laboratory assistants. The central laboratory and the regional laboratories have
separate staff to perform clinical chemistry tests whereas all the district laboratories
perform tests on rotational basis.
Table 5.5 Category of Staff in the laboratories
Laboratory No.
Pathologist
MT
Clin Lab Asst.
Clin. Chem. test
Total
1
0
0
2
Rotation
2
2
0
0
2
Rotation
2
3
0
0
4
Rotation
4
4
0
0
2
Rotation
2
5
0
1
9
2
10
6
4
3
46
7
60
7
0
0
2
Rotation
2
8
1
2
9
1
11
9
0
0
4
Rotation
4
10
0
0
3
Rotation
3
11
0
0
6
1
6
12
0
0
3
Rotation
3
13
0
0
2
Rotation
2
14
0
0
3
Rotation
3
15
0
0
3
Rotation
3
16
0
0
3
Rotation
3
17
0
0
2
Rotation
2
18
0
0
2
Rotation
2
19
0
0
3
Rotation
3
Total
5
6
116
11
127
5.1.4 IQC and EQA activities in the clinical chemistry laboratory in
Bhutan
The IQC and EQA activities are summarized in table 5.6. Only one of the
laboratories participates in EQAS but IQC is performed in all the laboratories by using
the commercial control materials. Only two laboratory personnel out of 127 staff have
been trained in basic laboratory quality assurance which indicated the needs of
participants to educate before implementing the EQAS. Accordingly three days
workshop was planned and trained on basic laboratory quality assurance. After the
training of the participants in workshop, comparatively better performance was seen in
post test than pretest. The result of high score (70-100%) was significantly increased
(figure 5.1). The knowledge on safety and precautions in handling the control
materials were adequate.
Table 5.6 IQC activities in the clinical chemistry laboratory in Bhutan
Laboratory Activities
Central
Regional
District
Remarks
laboratory
laboratory
laboratory
1.Participation in any EQAS
√
x
x
PPTC, Newzeland
2.
Use IQC Material
√
√
√
Commercial control
3.
Use lyophilized serum
√
√
√
*add 5ml DW
4.
Levels of IQC in use
2
3
2
Mimic patient
serum
5.
Prepare any IQC Material
x
x
x
Manpower shortage
6.
Use LJ chart
x
x
x
For IQC monitoring
7.
Calculate IQC range
x
x
x
Use Mfg‟s range
8.
Use volumetric pipette*
x
x
x
For reconstitution
9.
Wear gloves & gowns
√
√
√
Safety purpose
10. Disinfect vials &
√
√
√
Protect
samples
environment
Figure 5.1 The pre-test and post-test scores of the 27 laboratory staff participants
attended in 3 days workshop organized prior to the implementation of EQAC-DLB in
Bhutan. The figure shows rise in number of participants with high scores in 70-100%
in post-test compared to the pre-test.
5.2 Results of the study on characteristics of QC sample
Total of 1600 ml of serum was collected from 11 Bhutanese volunteers.
The pooled serum was divided into four parts to prepared 4 different levels as follows:
Level N
1
= 275 ml, Level N
2
= 310 ml, Level A
1
= 310 ml and Level A
2
= 310 ml.
5.2.1 Homogeneity test results for the EQA samples
At 99% confidence interval, no significant outliers were detected by
Cochran‟s test when compared with Cochran‟s critical values at 95% and 99%
confidence interval which were 0.727 and 0.838, respectively (53). Sampling variance
showed no significant difference from that of critical f values of inhomogeniety test
for all the analytes at 95% confidence interval with 6 degrees of freedom (table 5.7).
Table 5.7 Results of the homogeneity test for the EQA samples
5.2.2 Assigned values derived from the results of ISO 15189 certified
laboratories
The assigned values for our control materials are shown in the table 5.8.
There are two levels of normal sample designated as N
1
and N
2
and another two
levels for abnormal as A
1
and A
2
. The Robust mean and SDs were calculated using
ISO 15328 and the mean was used as assigned values.
Table 5.8 The mean of the results from three ISO 15189 certified laboratories in
Thailand designed as assigned values of control material in this project.
Mean and CV of assigned values for 4 levels
Test
N
1
N
2
A
1
A
2
s
SD
CV
SD
CV
SD
CV
SD
CV
Glu
77
1.1
1.43
69
1.2
1.7
117
2.0
1.7
136
1.5
1.1
BUN
11
0.38
3.45
10
0.7
7.0
19
0.75
0.3
22
0.6
2.7
Creat
0.9
0.08
8.9
0.83
0.07
8.4
2.1
0.09
4.3
2.5
0.12
4.8
AST
20
1.5
7.50
18
1.3
7.2
19
2.0
10.5
16
2.0
13
ALT
14
2.0
14.3
12
1.6
13.3
65
1.1
1.7
155
3.0
1.9
ALP
65
2.6
4.00
57
2.7
4.7
293
13
4.4
337
19
5.6
T.B
0.55
0.05
9.09
0.55
0.05
9.1
1.5
0.07
4.7
2.0
0.14
7.0
T.P
7.2
0.21
2.92
6.6
0.2
3.0
7.2
0.21
2.9
7.2
0.19
2.6
Alb
4.6
0.07
1.52
4.2
0.07
1.7
4.5
0.10
2.2
4.6
0.08
1.8
5.2.3 Results of the stability tests
As described in ©2006 IUPAC Protocol, at 95% confidence interval, the
results of the sample A
1
and N
1
stored at 2-8
o
C and -20
o
C were statistically compared
with initial values as shown in table 5.9-5.12. Line graphs are plotted for the means of
the duplicate results along with initial values (figure 5.2 and 5.3).
Table 5.9 Results of the stability test for the sample level A
1
stored at 2-8
o
C.
Analytes
Glu
BUN
Creat
AST
ALT
ALP
T.B
T.p
Alb
Initial values
114
19
2.1
19
64
282
1.5
7.0
4.5
286.1
Mean
109.5
18.21
2.17
20.00
54.82
1
1.41
7.76
4.49
SD
1.05
0.51
0.04
1.16
4.86
10.87
0.06
0.90
0.06
CV
0.96
2.80
1.93
5.79
8.87
3.80
4.44
11.6
1.42
P-values
0.00
0.04
0.41
0.16
0.00
0.30
0.03
0.03
0.90
Table 5.10 Results of the stability test for the sample A
1
stored at -20
o
C
Analytes
Glu
BUN
Creat
AST
ALT
ALP
T.B
T.p
Alb
Initial values
114
19
2.1
19
64
282
1.5
7.0
4.5
Mean
110.9
18.60
2.15
19.10
60.30
290.00
1.49
7.74
4.48
SD
1.14
0.65
0.05
1.98
1.64
13.31
0.10
0.96
0.06
CV
1.03
3.50
2.13
10.37
2.72
4.59
6.49
12.44
1.30
P-value*
0.70
0.40
0.20
0.90
0.70
0.16
0.25
0.04
0.80
Table 5.11 Stability results of EQA sample N
1
stored at 2-8
o
C
Analytes
Glu
BUN
creat
AST
ALT
ALP
T.B
T.p
Alb
Initial values
76
11
0.91
21
14
63
0.6
7.1
4.5
Mean
71.3
11.3
0.99
20.3
10.4
66.9
0.53
7.8
4.47
SD
1.60
0.27
0.02
1.89
1.47
5.67
0.02
0.94
0.06
CV
2.24
2.39
2.02
9.31
14.13
8.48
3.77
12.05
1.34
P-value
0.00
0.68
0.00
0.40
0.00
0.13
0.03
0.03
0.17
Table 5.12 Stability results of EQA sample N
1
stored at -20
o
C
Analytes
Glucose
BUN
Creat
AST
ALT
ALP
T.B
T.p
Alb
Initial values
76
11
0.91
21
14
63
0.6
7.1
4.5
Mean
74.30
11.10
0.95
22.20
13.40
69.60
0.55
7.84
4.51
SD
2.17
0.42
0.04
3.23
1.08
7.59
0.02
0.99
0.05
CV
2.92
3.77
4.30
14.56
8.09
10.91
4.16
12.64
1.22
P-value
0.08
0.51
0.08
0.39
0.50
0.03
0.4
0.03
0.8
(a)
(b)
(c) (d)
(e) (f)
Figure 5.2 Graphs 52(a) to 52(i) show the distribution of the mean of the duplicate
results of glucose, BUN, creatinine, AST, ALT, ALP, total bilirubin, total protein and
albumin in sample level A
1
stored at 2-8
o
C and -20
o
C analyzed in 3, 5, 7, 8 and 9
months.
(g)
(h)
(i)
Figure 5.2 Graphs 52(a) to 52(i) show the distribution of the mean of the duplicate
results of glucose, BUN, creatinine, AST, ALT, ALP, total bilirubin, total protein and
albumin in sample level A
1
stored at 2-8
o
C and -20
o
C analyzed in 3, 5, 7, 8 and 9
months (continued).
(a) (b)
(c) (d)
(e) (f)
Figure 5.3 Graphs 5.3(a) to 5.3(i) show the distribution of the mean of the duplicate
results of glucose, BUN, creatinine, AST, ALT, ALP, total bilirubin, total protein and
albumin in sample level N
1
stored at 2-8
o
C and -20
o
C analyzed in 3, 5, 7, 8 and 9
months.5.
(g) (h)
(i)
Figure 5.3 Graphs 5.3(a) to 5.3(i) show the distribution of the mean of the duplicate
results of glucose, BUN, creatinine, AST, ALT, ALP, total bilirubin, total protein and
albumin in sample level N
1
stored at 2-8
o
C and -20
o
C analyzed in 3, 5, 7, 8 and 9
months (continued).
5.3 Results of the assessment of interlaboratory precision by
comparing average monthly CV for each analyte with that of CCV
from CLIA 88’
The average CV between 2 trials of samples was calculated using the
overall mean and SD. The CVs were compared with CCV from CLIA PT criteria
(figure 5.4). The CVs of all analytes were seen higher than the CLIA criteria during
the first round which showed improvement towards the end of the study except for
total bilirubin and total protein.
Figure 5.4 Average CV of the results of two trials in the beginning and at the end of
the program compared with CCV from CLIA PT Criteria.
5.4 Performance assessment using MVIS
5.4.1 Assessment on instrument and method performance using MVIS
Mean variance index scores (MVIS) for each analyte and each level was calculated by
grouping the results based on the sample levels, methods and instruments used by the
laboratories (figure5.5-5.13).
Figure 5.5 Methods and instruments performance on glucose: photometers using the
GOD/POD method and Hitachi 911 using GOD/PAP method shows MVIS<200 but
Smartlab shows MVIS>200 for N
1
and N
2
level.
Figure 5.6 Methods and instruments performance on BUN: photometers using the
Bertholet method (endpoint) shows MVIS>200 for sample N
2
and A
1
. Whereas two
autoanlyzers using kinetic method shows MVIS<150.
Figure 5.7 Methods and instruments performance on creatinine: Both the photometers
and Hitachi 911 using the same methods show MVIS<150 and Beckman CX PRO
using the Rosaliki method also shows MVIS<150 except for N1.
Figure 5.8 Methods and instruments performance on AST: the two photometers and
Hitachi 911 using different methods show MVIS<200 and Beckman CX PRO shows
MVIS<100.
Figure 5.9 Methods and instruments performance on ALT: the two photometers using
IFCC methods show MVIS<200. Hitachi 911 and Beckman CX PRO using different
methods shows MVIS <100 except for N
1
.
Figure 5.10 Methods and instruments performance on ALP: the two photometers
using pNPP methods show MVIS<150 for A
1
& A
2
but > 200 for N
1
& N
2
. Hitachi
911 and Beckman CX PRO using different methods show MVIS <150.
Figure 5.11 Methods and instruments performance on total bilirubin: the two
photometers and Beckman CX PRO using Modified Jendrassik method show
MVIS≤150 for A1 & A
2
but > 200 for N
1
whereas Hitachi 911 using DCA method
shows MVIS<150.
Figure5.12 Methods and instruments performance on total protein: all the analyzers
use Biuret methods. The BA-88 shows better performance of the two photometers
with MVIS≤200. Two autoanlyzers show MVIS<100 for all the four levels.
Figure 5.13 Methods and instruments performance on albumin: BA-88 and Hitachi
911 using the same methods show MVIS<200 but SM-3000 shows MVIS >200 for all
the levels and Beckman CX PRO using the Bromocresol purple method shows
MVIS<200 for all the four levels.
5.4.2 Assessment of the performance on four concentration levels of
samples
In order to assess the performance on the four levels of samples, the
monthly VIS were divided into four groups based on the concentration levels which
were designated as N
1
, N
2
, A
1
and A
2
(figure 5.14). The performance is observed
better in higher concentration of the analytes i.e. A
1
and A
2
which shows VIS < 200
except for total protein in A
2
. VIS > 200 was observed for ALP and total bilirubin in
level N
1
and VIS>200 for BUN, ALP, total bilirubin and total protein in level N
2
.
Thus high VIS is observed for each analyte in decreasing order of their concentration
levels.
Figure 5.14 Mean VIS of each analytes were compared among four levels of the
control material. The performance is observed better in higher concentration of the
analytes i.e. A
1
and A
2.
5.4.3 Laboratory performance categorized based on VIS.
The overall performance evaluation by grouping VIS into three categories
A, B and C. Interpretation: A= VIS < 200 as acceptable; B = VIS: 200-250 Needs
improvement; C = VIS>250 as unacceptable. The acceptable VIS as „A‟ are 63%,
60%, 66% 69%, 73% and 74% , 75%, 76%, 79% and 68% for November to August,
respectively. Unacceptable VIS as „C‟ are 27%, 31%, 25%, 21%, 20%, 20%, 20%,
and 14% in November 2009 to July 2010, respectively. The bar graph of the overall
mean variance index scores using VIS of the all the laboratories were plotted to assess
performance in each month. As shown in the figure 5.15, the performance is seen
gradually improved but abruptly declined in the last round due to reagent drift. The
figure 5.16 shows high OMVIS in December and August in which sample level N
1
and A
2
were analyzed.
Figure 5.15 The overall performance evaluation by grouping VIS into three categories
A, B and C. Interpretation: A= VIS < 200 as acceptable; B = VIS: 200-250 Needs
from November 2009 till August 2010.
Figure 5.16 OMVIS for each month of the study. Figure shows the highest VIS in the
month of December, 2009 and august 2010 in which the samples A
2
and N
2
were
analyzed.
5.5 Modified Cpk as performance indicator
The mCpk was calculated for glucose to study its application in evaluation
of the laboratory performance (figure 5.17). The negative mCpk reflects adequate
precision but bias higher than total allowable error. Three laboratories have scored
negative mCpk and not even single laboratory scored mCpk>200. The advantage of
the mCpk over Cpk is that mCpk is in numerical value which is easy for reading and
interpretation unlike the Cpk which is in decimals (figure 5.18). Another advantage is
its ability to reflect both accuracy and precision unlike VIS which shows performance
only in terms of bias. As in figure 5.19, the Lab 1410 shows the highest MVIS which
means the laboratory has the high bias. Correspondingly, the mCpk from the same
laboratory shows very low mCpk which indicates that, the laboratory has high bias
and high imprecision because the mCpk is effected by bias and CV.
Figure 5.17 mCpk for glucose among 19 laboratories. Four laboratories are seen with
mCpk in negative magnitude indicating the presence of high bias.
Figure 5.18 Cpk for glucose among 19 laboratories. The performance pattern is seen
exactly similar to mCpk but Cpk is in decimal points.
Figure 5.19 MVIS of glucose for each laboratory. Lab 1410 shows the highest MVIS
which indicates bias in glucose measurement. But corresponding Cpk values shows
the presence of imprecision.
5.6 EQAS performance in relation to the IQC
The IQC results from seven laboratories are in their IQC range in both
glucose and AST (table 5.13). However, 2 laboratories show slightly unacceptable CV
in either of the concentration levels (lab 1405 and 1412). The EQA results correspond
with that of IQC in terms of the range and precision. The CVs for AST in 4
laboratories show imprecision with indication of calibration problem in lower
concentration. One of the laboratories (Lab 1418) shows similar problems in higher
concentration. The EQA results for AST corresponds with that of IQC in terms of the
range and precision except for two laboratories in which decreased values have been
reported possibly due to random errors. The lab-1418 showed the highest imprecision
in glucose and Lab-1404 in AST (table 5.13).
Table 5.13 The IQC and EQAS results for glucose and AST from 7 laboratories
to assess the laboratory performance in relation to IQC.
Abbreviation and interpretation: T = Trial; √ = in range; ï = imprecision with linearity problem; c =
calibration error; A: IQC results compared with IQC range; B: = interpretation of laboratory precision
(CV). C = EQA results interpretation; improvement; R = random error; A: IQC results compared with
IQC range at mean ±2SD; B: = interpretation of laboratory precision (CV). C: interpretation of the
EQAS results compared with IQC and EQAS range at mean ±3SD. T
1
, T
13
and T
18
are the IQC and
EQA results of level N
1
analyzed on the same days. Similarly, T
2
, T
14
and T
17
are for level A
1
.
Figure 5.20 OMVIS for each laboratory in Bhutan. The Lab-1418 which shows the
highest OMVIS was chosen to monitor the performance during the entire period of
study. The IQC and EQAS results for glucose and AST from 7 laboratories to assess
the laboratory performance
5.7 Analysis of performance characteristics of laboratory with highest
OMVIS in EQAS and high imprecision in IQC
Data from lab-18 in table 5.13 that shows the high imprecision in both
high and normal level and the highest OMIVIS was chosen to analyze the
performance characteristics on glucose in EQAC-DLB for the entire period of the
study. The performance showed gradual improvement indicated by decreasing VIS
from November till March. There is sudden rise in MVIS in April 2010 with reversion
in the subsequent rounds. For the AST, the pattern shows inconsistent performance in
the beginning and gradually improved until June 2010. The MVIS for both glucose
and AST increased in August probably due to change in reagents.
Figure 5.21 The MVIS of glucose for the Lab-1418 during the entire period of
implementing EQAC-DLB.
Figure 5.23 The MVIS of AST for the lab 1418 during the entire period of
implementing EQAC-DLB.
CHAPTER VI
DISCUSSION
Clinical laboratory service is an indispensible part of health services. In
order to ensure reliable results from the laboratory, it is very necessary to have the
quality assurance program in all the laboratories. Although the quality assurance
program has not been introduced yet in Bhutan, QA practices such as daily QC has
been performed in almost all the laboratories. However, with increasing number of
patients in demand for better quality of health services, it‟s time to establish a
sustainable QA program in Bhutan so as to enable to monitor the laboratory testing.
This is possible by establishing an EQAS as it is an effective tool for assessing and
improving the quality of laboratory performance.
Therefore, we implemented EQAS in clinical chemistry for district
laboratories in Bhutan for the period of ten months which began with collection of
information from the laboratories. In our preliminary data analysis, use of micropipette
for reconstituting control material; inadequate knowledge in using IQC range and
Levey-Jennings chart; use of manufacturer‟s range as IQC range and only two
laboratories personnel having trained in basic laboratory QA indicated the current
situation of the IQC activities in Bhutan. These evidence of drawbacks showed that,
the participants need to educate prior to the implementation of EQAS. Accordingly,
three days workshop was conducted to brief on scope and importance of EQAS with
inclusion of basic lessons on IQC. After training, the results of the pre-test and post-
tests showed that our workshop was effective for the participants. However, only 27
participants out of 77 laboratory personnel performing clinical chemistry attended the
workshop. Most of the participants attending the workshop were in-charge of the
laboratories. Rest of their colleagues could not attend this workshop due to the
shortage of manpower as remaining staff had to perform the routine tests. As part of
the activities in the future, all the laboratory personnel performing clinical chemistry
are recommendable to be included in such workshop. Bhutan being the landlocked
country, the hospitals are located at various regions separated by rivers, valleys and
mountains. Moreover, winding roads are often washed out by landslides and
communication facilities are disrupted causing difficulties in transportation. Thus,
lyophilized serum is only the suitable control material for EQAC-DLB in Bhutan.
Homogeneity test was performed to check for vial-to-vial variation in our
homemade lyophilized human serum control material. Through our careful sample
processing and the preparation technique, no inhomogeneity was oberserved in our
sample.
We carried out stability tests for the sample stored at 3 chosen conditions
to study the stability of our sample during the entire period of study. A few analytes in
our homemade EQA sample showed time and temperature dependent changes over the
period of studies. There are many literatures describing the instability of the biological
compounds in lyophilized QC serum, reconstituted QC serum and serum in whole
blood stored at various temperatures (61-64). Since we prepared our control material
without preservatives and stabilizers, the slight decrease in glucose levels could be due
to degradation by microbial agents during the storage (61, 64). Increase in creatinine
may be attributed to the formation of pseudocreatinine in Jaffe‟s reaction.
Decrease in ALT level could be due to the loss of enzyme activities in prolonging
storage and the interference by LDH on ALT may be considered (62, 63). The
increase in ALP activity has been reported as due to either release of alkaline
phosphatase from a complex containing the inactive or partially inactive enzyme or by
the dilution effect of inhibitors during reconstitution (65). Gradual decrease in total
bilirubin could be due to photo degradation of bilirubin brought about by prolonging
exposure to the light outside after reconstitution (62, 63). The increase in protein
levels could be attributed to the liberation of free proteins from the glycoprotein and
increase bacterial activities would have produced some enzymes and other microbial
products that would contribute in increasing the protein levels (61-63).
The variety of statistical approaches used to assess stability also affects the
stability results. It has been reported that about 5% of the results are expected to show
instability due the choice of the statistics (50). The studies have been done by Özbek
et al. (64) who also prepared lyophilized human serum from Turkish people without
any additives and stabilizers. They concluded that their samples were suitable up to 5
months and about 39% of the glucose was lost at the end of 5 months and the storage
temperature hasn‟t been provided. In our studies, the glucose loss was only 5% at 2-
8
o
C and 3% at -20
o
C at the end of 9 months. CVs for AST and ALT for higher
concentration (A
1
) were greater than that of low concentration (N
1
) which could be
due to concentration dependent calibration errors. Constant CV of 12% was seen for
total protein which may be due to the time and temperature dependent increase during
storage condition.
Through a constant feedback with suggestions for action in case of
variations in the results, there has been gradual improvement in the performance.
Although inconsistent, the gradual decrease in overall CV compared with CLIA PT
criteria showed improvement of the interlaboratory variation. It is also a reflection of
the decrease incidence of random errors committed by individual laboratories.
However, the trends looked wavy and CVs were inconsistent depending upon the
concentration of the samples. The percentage results in acceptable VIS (i.e. VIS ≤
200) was increased accompanied by decrease percentage of unacceptable VIS (VIS >
250) implying that performance has been improved over the period of ten months
through corrective action following our recommendation except in August.
Many factors needed to be considered for the laboratories showing
inconsistent performance and the exact source of laboratory errors could not be
defined. Some common causes of the random and systemic errors need to be rule out
which can be detected by the routine QC procedures. The common sources of random
errors in reconstitution are use of inappropriate pipettes; errors in pipetting technique,
insufficient time for complete dissolving of lyophilisate and inappropriate mixing
technique. As the general requirements of the laboratory procedures, the laboratory
personnel should ensure that, there is no mismatch of the reagent-instrument
specification; calibration errors of the analyzers; use of expired reagents and mismatch
between methods in control and reagents. Prolong standing of reconstituted control at
room temperature causes time and temperature dependent increase activities of ALP
(65). These factors would have lead to unfair assessment of those laboratories using
semiautomatic methods which needs longer time in incubation, reagent preparation
and manual sequential running of the sample solution for analysis.
In assessment of the method and instrument performance, the laboratories
using Hitachi 911 and Smart Lab with GOD/PAP method for glucose generated higher
VIS for glucose than the photometers. Beckman CX5 user reported to have problem in
calibration for glucose and uses semiautomatic analyzer. The photometer users using
endpoint methods yielded better results than automated analyzers for glucose. This
effect could be indication of a method and instrument associated biases which need to
be defined clearly through careful inspection of the analyzers. Ziegenhor et al. (66)
outlined the advantages of using kinetic method in glucose determination who
concluded that the method is more accurate, rapid and better precision than the end
point method.
Mitchell (67) has described the advantages and purpose of the automation
in the clinical chemistry. Increase efficiency of laboratory operations; better accuracy
and precision of the results; high sample throughput; rapid sample analysis, automated
data analysis avoiding clerical errors; and solving the staff shortages are some
advantages of the automation in laboratory medicine. But an unusual performance by
the laboratories using automated Hitachi 911 and Smart Lab could be due to
preanalytical factors and poor inspection of the analyzer. The similar concentration
dependent increase in MVIS is observed in performance on BUN using Bertholet
method in BA-88 group. The laboratory using Hitachi 911 with Urease GLDH kinetic
method and Beckman CX5 using UV kinetic method for BUN yielded better results
than both the correlating with Mitchell‟s study on trends of automation. In creatinine
determination, the users of the photometer and Hitachi 911 using Jaffe‟s method and
Rosalinik method for Beckman CX5 produced VIS of ≤ 150 which shows good
performance for the creatinine but shows pronounced concentration dependent high
MVIS by Hitachi 911 and Beckman CX5. For the AST/ALT, Beckman CX5 operated
using Henry method Performed comparatively better for these enzymes with MVIS <
100. In overall, photometers using modified IFCC method and Hitachi 911 using
optimized IFCC UV kinetic method yielded MVIS <200 showing acceptable
performance. In determination of ALP all the 3 instrument yielded acceptable MVIS
except for N
1
and N
2
by two photometers (pNPP method) which produced MVIS >
200 indicating a concentration dependent errors. High MVIS was produced for total
bilirubin, total protein and albumin by the photometer users with Jendrassik, Biuret
and BCG methods, respectively.
The higher MVISs were observed in A
2
and N
2
. It is probably due to the
exaggeration of the MVIS in December 2009 and August 2010 as participants
analyzed sample level A
2
and N
2
which produced the highest MVIS during that two
months. Sudden increase in MIVIS in December 2009 could be due to the change in
reconstitution technique as participants were supplied with volumetric pipette and
stopped using micropipette following our suggestions and recommendations. Volume
errors would have been committed in the initial phase of the changing technique.
The comparison of our EQA results and IQC showed good correlation in
terms of range and precision which indicated that our program not only improved the
performance on our samples but also improved the IQC of the participating
laboratories. However, the IQC ranges for most of the laboratories were too wide with
disparity in performance on two concentration levels. Disparity characterized by
increase or decrease CV in either for the high or normal control indicates the
concentration dependent calibration errors in the laboratories. The high CVs in both
high and normal levels accompanied by increase IQC range compared to
manufacture‟s range indicated imprecision probably due to more incidence of random
errors. Therefore, the better performance on high concentration compared to the lower
concentration level indicated the errors due to concentration dependent calibrations.
Some of the participating laboratories reported to have received reagents towards the
end of our study which might have lead to method associated variations causing
elevated VIS in the last rounds.
VIS system has been used in many international EQAS but it indicates
more of bias of the individual results from the target values. Since we use CCV as the
constant scaling factor, VIS system is not very sensitive to indicate the precision and
reproducibility. Thus our mCpk is a useful scoring index to be applied as alternative
scoring index to evaluate performance of the laboratory which can reflect both bias
and precision. It is particularly useful to apply in EQAS end-of-cycle summary data
which can reflect the laboratory precision and effect of the change in reagent and
control lots. It is applicable especially for those laboratories using manual methods
where random errors resulted from the preanalytical factors are quite common. The
preanalytical factors generate results with high imprecision and ultimately high bias.
Such high imprecision yields high group SD thus resulting in very low capability
index with decimal points. Therefore, in this study, due to the high imprecision and
high bias from the participants, the mCpk generated were incredible. Although the
laboratory performance is improved, the source of errors could not be defined for the
laboratories committing persistent errors.
The experiment on calibration, methods and instrument verification may
be performed as part of the immediate activities through careful inspection of the
analyzers and working system in each laboratory in Bhutan. Inability to perform
glucose by Beckman CX5 needs to be addressed and find out the causes and solutions.
The high glucose values generated by Smartlab shows that the instrument selection
needs to be done by the personnel experienced in verifying the instrument
performance. The constant high glucose value by Hitachi 911 requires careful
inspection to rule out some systemic errors to ensure reliable results for the patients.
Although, the efforts and plans are already in place to improve the quality of
laboratory services, there is need to define the quality goal and outline the mission for
the next few years to establish a sustainable clinical laboratory quality assurance
system in Bhutan.
To ensure sustainability of laboratory QA system, team has to be formed
including national team for QA management; team to ensure undisrupted supply of
reagents; training & education and analytical inspection (setting target ranges, result
interpretation & reporting) to improve the overall medical care for the patients. Beside
this, proper guidance is required for laboratory personnel with more emphasis on
quality control. There is need to revise the curriculum for the current laboratory
technicians at the Royal Institute of Health Science by incorporating lessons on basic
laboratory quality assurance.
Although we planned our study for twelve months, we operated up to 10
months for logistic reasons. Out of 19 participants in the beginning, one of the
laboratories discontinued its participation who reported as machine failure for the
cause of discontinuation. Number of results being reported by participants was
decreasing every month due to the shortage of reagents. These limitations faced by our
participants showed that the lack of support for the equipment maintenance, shortage
of reagents and quality control materials has seriously undermined the ability to
perform laboratory testing in Bhutan. Therefore, the inconsistent and incomplete
results due to erratic supply of reagents at the district laboratories; the high expenses
on communication with the participants and coordinators; insufficient time for the
investigation were the major challenges in this study.
CHPATER VII
CONCLUSION
The primary objective of any EQA scheme in laboratory medicine is to
support the improvement of the quality of laboratory service. It is achievable only
through evaluation of analytical performance followed by continuous education and
training in required field. In line with this concept of EQAS, from the results of our 10
months program (EQAC-DLB), we realized the importance of EQAS which played
crucial role in monitoring the laboratory performance in Bhutan. Its role in laboratory
medicine will be paramount if the QC materials are prepared within organization or
within the country.
Through our recommendations along with our feedbacks, laboratory
performance has been improved both in terms of IQC and EQAS as shown by the
results of the assessment. Therefore, like any other EQA program, EQAC-DLB was
very effective not only in monitoring the quality of laboratory performance but also
provided educational opportunities for the laboratory personnel. Our results showed
that, the establishment of national EQAS by preparing homemade lyophilized human
serum is effective and useful scheme in terms of resources and time to monitor the
laboratory performance in Bhutan. The data generated in this study demonstrated the
advantages of mCpk over VIS and Cpk to be used as scoring index in EQAS.
However, the cause of inconsistent performance by some of the
laboratories and the source of concentration dependent errors need further
investigation. Further training is required for all the laboratory personnel to upgrade
and update the knowledge in laboratory quality assurance. Extended study for a period
of 1-2 years may be carried out after establishing quality assurance program in Bhutan
to study the effect and efficiency of EQAS in Bhutan and mCpk may be used to
compare the laboratory performance over time.
REFERENCES
1. Ministry of Health in Bhutan. Annual health bulletin 2008: Century of progress in
health. [Online]. 2008 [cited 2009 May 20]; Available from:
URL:http://www.health.gov.bt/statistics/bulletins/ahb2008/healthFacilitie.
pdf
2. Ministry of Health in Bhutan. Report of the national health survey in Bhutan.
[Online]. 2000 [cited 2010 Oct 25]. Available from:
URL: http://www.health.gov.bt/healthsurvey2000.pdf
3. Ministry of Health in Bhutan. Quality assurance and standardization. [Online]. 2008
[cited 2010 Sept 20]; Available from:
URL: http://www.health.gov.bt/qasd.php
4. Victor WL. Laboratory services in Bhutan: An evaluation and recommendation.
Ministry of Health;2004.
5. Quality Assurance Standardization Division. Clinical laboratory standards in
Bhutan. 1
st
ed. Thimphu; 2008.
6. World Health Organization. Establishment of quality systems and accreditation in
health laboratories-report of an inter-country workshop. 2007 [Online].
[cited 2010 Oct 23]. Available from:
URL:http://www.searo.who.int/LinkFiles/BCT_Reports_HLM- 393.pdf
7. Deom A, Aouad RE, Heuck CC, Kumari S, Lewis SM, Uldall A et al. World
Health Organization: requirements and guidance for external quality
assessment schemes for the health laboratory. [Online]. 1999 [cited 2010
Oct 20];Available from:
URL:http://whqlibdoc.who.int/hq/1999/WHO_DIL_LAB_99.1_REV.1.
pdf
8. Burnett L, Hegedus G, Chesher D, Burnett J, Costaganna G. Application of process
capability indices to quality control in a clinical chemistry laboratory. Clin
Chem 1996;42(12):2035-7.
9. Bais R. Use of Capability Index to improve laboratory analytical performance. Clin
Biochem Rev 2008;29(Suppl 1):S27-31.
10. ISO 15189. Medical laboratories-particular requirements for quality and
competence. 2
nd
ed. Geneva: International Organization for Standard;
2007.
11. Sheiner LB, Wheeler LA, Moore JK. The performance of delta check methods.
Clin Chem 1979;25(12):2034-7.
12. Abbott diagnostics: X-B (Moving Averages) - Cell-DYN® 3000 Series Systems.
[Online]. 2002 [cited 2010 June 2]; Available from:
URL:http://www.abbottdiagnostics.co.nz/viewFile.cfm?file=x_b_moving_
avgs_ P1_1_90.pdf
13. Henry RJ. Use of the control chart in clinical chemistry. Clin Chem
1959;5(4):309-19.
14. Westgard JO, Barry PL, Hunt MR, Groth T. A multi-rule Shewhart chart for
quality control in clinical chemistry. Clin Chem 1981;27(3):493-501.
15. Bower KM. Using Cusum charts to detect small process shift. [Online]. 2002
[cited 2010 Oct 2]; Available from:
URL:http://www.minitab.com/uploadedFiles/Shared_Resources/Documen
ts/Articles/cumulative_sum_charts_for_small_shifts.pdf
16. Cumulative sum chart (CUSUM). [Online]. 2002 [cited 2010 June 2]; Available
from: URL:http://www.variation.com/cpa/help/hs108.htm
17. Libeer JC, Baadenhuijsen H, Fraser CG, Petersen PH, Ricos C, Stockl D et al.
Characterization and classification of external quality assessment schemes
(EQA) according to objectives such as evaluation of method and
participant bias and standard deviation. External Quality Assessment
(EQA) Working group on analytical goals in laboratory medicine. Eur J
Clin Chem Clin Biochem 1996;34(8):665-78.
18. Sciacovelli L, Secchiero S, Zardo L, Plebani M. External quality assessment
schemes: need for recognised requirements. Clin Chim Acta 2001
20;309(2):183-99.
19. Hill P, Uldall A, Wilding P. Guidelines for improving analytical quality by
establishing and managing EQA schemes: examples from basic chemistry
using limited resources. 1996 [Online]. [cited 2010 Aug 10]. Available
from: URL:http://www.ifcc.org/PDF/Fundamentals-for-EQA.pdf.
20. Sciacovelli L, Zardo L, Secchiero S, Plebani M. Quality specifications in EQA
schemes: from theory to practice. Clin Chim Acta 2004;346:87-97.
21. RCPA-QAP. Reflections on 20 Years of the QAP. [Online]. 2010 [cited 2010 May
27]. Available from:
URL:http://www.rcpaqap.com.au/chempath/qaphistory.cfm
22. Bais R. EQA from an Australian perspective. Clin Biochem Rev 2007;28(4):175-
8.
23. RCPA Quality Assurance Program Pty Ltd. Target value assignment methods.
[Online]. 2010 [cited 2010 Sept 20]. Available from:
URL:http://www.rcpaqap.com.au/chempath/targetvalues.pdf
24. United Kingdom national external quality assessment service. [Online]. 2010
[cited 2010 Oct 15]. Available from:
URL:http://www.ukneqas.org.uk/content/Pageserver.asp
25. WEQAL. International external quality assessment schemes. [Online]. 2010 [cited
2010 Aug 27]. Available from: URL:http://www.ieqas.org.uk/weqal.htm
26. Bullock DG. External quality assessment schemes for clinical chemistry in the
United Kingdom. Ann 1
st
Super Sanita 1995;31(1):61-9.
27. Whitehead TP, Woodford FP. External quality assessment of clinical laboratories
in the United Kingdom. J Clin Pathol 1981;34(9):947-57.
28. Randox Laboratories. The future of healthcare. [Online]. 2010 [Cited 2010 May
25]; Available from: http://www.pathologyinpractice.com.
29. Randox Laboratories. RIQAS explained the largest global EQA scheme in the
world. [Online]. 2010 [cited 2010 Aug 17]; Available from:
URL:http://www.randox.com/brochures/PDF%20Brochure/LT033.pdf
30. Randox international external quality assessment schemes. [Online]. 2010 [cited
2010 Aug 25]; Available from: URL:http://www.riqas.com/EQA.asp
31. Historical perspectives of CDC. Online]. 2010 [cited 2010 Aug 29]. Available
from: URL:http://www.cdc.gov/mmwr/preview/mmwrhtml/00042732.htm
32. American Association for Clinical Chemistry. August 2008 clinical laboratory
news: CDC cholesterol standardization program celebrates 50 years of
service. Online]. [cited 2010 August 20]; Available from:
URL:http://www.aacc.org/publications/cln/2008/august/Pages/inside_080
8.aspx
33. CDC. External QA programs for clinical laboratories. [Online]. 2010 [cited Sept
25, 2010]; Available from:
URL:http://wwwn.cdc.gov/mlp/EQA_qlist.aspx
34. Lawson NS. Quality assurance program in the United States. Ann 1
st
super Sanita
1995;31(1):21-35.
45. American Association for Clinical Chemistry. [Online]. 2010 [cited 2010 Sept 23];
Available from: URL:http://www.aacc.org/about/Pages/default.aspx
35. College of American Pathologist. Laboratory improvement program [Online].
2010 [cited 2010 Sept 29]; Available from: URL:http://www.cap.org
36. Bio-Rad Laboratories. [Online]. 2010 [cited 2010 Oct 10]; Available from:
URL:http://en.wikipedia.org/wiki/Bio-Rad_Laboratories
37. BioRad laboratories. An effective tool for evaluating analytical performance
[Online]. 2010 [cited 2010 Oct 15]; Available from:
http://qcnet.com/Portals/50/PDFs/UnityInterlab-Booklet2007.pdf.
38. BioRad laboratories. Bio-Rad external quality assurance services. [Online].
2010 [cited 2010 Oct 2]; Available from:
URL:http://qcnet.com/EQAS/tabid/334/Default.aspx
39. Promptmas C, Prijavudhi A, Pavaro U. Quality assessment in clinical chemistry: a
Thailand‟s experience. Southeast Asian J Trop Med Public Health
1999;30(Suppl 3):46-9.
40. The external quality assessment schemes in Thailand. Rinsho Byori
2002;50(2):121-541.
41. Selvakumar R, Swaminathan S, Fleming J. Report of a trial with lyophilized QC
sera in the CMCH/ACBI EQAS Programmes in India. Indian J Clin
Biochem 2004;19(1):24-33.
42. Geethanjalai F.S, Fleming J, Swaminathan S, Selvakumar R. External quality
assurance-role of ACBI/CMC Scheme. Indian J Clin Biochem
2006;21(1):211-12.
43. Christian Medical College, Vellore, India. External quality assurance scheme.
[Online]. 2010 [cited 2010 May 30]. Available from:
URL:http://home.cmcvellore.ac.in/clinqc/aboutEQAS.aspx
44. Bretaudlere JP, Dumont G, ReJ R, BaIlly M. Suitability of control materials.
General principles and methods of investigation. Clin Chem 1981;27(6):798-805.
45. Browning DM, Hill PG, Olazabal DAV. Preparation of stabilized liquid quality
Control serum to be used in clinical chemistry (WHO Lab/86.4). Geneva:
WHO;1986.
46. Kenny AP, Eaton RH. Practical guidelines for the preparation of quality control
sera for use in clinical chemistry (WHO Lab/81.4). Geneva: WHO;1981.
47. Eckfeldt JH, Copeland KR. Accuracy verification and identification of matrix
effects. The College of American Pathologists' Protocol. Arch Pathol Lab
Med 1993;117(4):381-6.
48. Henriksen GM, Pedersen MM, Norgaard I, Blom M, Blou L, Blaabjerg O, et al.
Minimally processed fresh frozen human reference sera: preparation,
testing, and application to international external quality assurance. Scand J
Clin Lab Invest 2004;64(4):293-308.
49. Ferrero CA, Carobene A, Ceriotti F, Modenese A, Arcelloni C. Behavior of frozen
serum pools and lyophilized sera in an external quality-assessment
scheme. Clin Chem 1995;41(4):575-80.
50. Hartmann AE, Juel RD, Barnett RN. Long-term stability of a stabilized liquid
quality control serum. Clin Chem 1981;27(8):1448-52.
51. Wilson SS, Guillan RA, Hocker EV. Studies of the stability of 18 chemical
constituents of human serum. Clin Chem 1972;18(12):1498-503.
52. Elin RJ, Gray BA. Liquid and lyophilized quality-control materials compared for
use in continuous-flow analysis. Clin Chem 1984;30(1):129-31.
53. Thompson M, Ellison SLR, Wood R. The international harmonized protocol for
the proficiency testing of analytical chemistry. Pure Appl Chem
2006;78(1):145-96.
54. ISO Guide 35. Reference materials-general and statistical principles for
certification. Geneva: International Organization for Standard; 2006.
55. ISO 13528. Statistical methods for use in proficiency testing by interlaboratory
comparisons. 1
st
ed. Geneva: International Organization for Standard;
2005.
56. ISO/IEC 43-1. Proficiency testing by interlaboratory comparisons: Part 1.
Development and operation of proficiency testing schemes. 2
nd
ed.
International Organization for Standard & International Electrotechnical
Commission. Geneva; 1997.
57. Ricos C, Baadenhuijsen H, Libeer CJ, Petersen PH, Stockl D, Thienpont L, et al.
External quality assessment: currently used criteria for evaluating
performance in European countries, and criteria for future harmonization.
Eur J Clin Chem Clin Biochem 1996;34(2):159-65.
58. Rhoads DG. Establishing performance standards: a practical approach. [Online].
2010 [cited 2010 June 18]; available from:
URL:http://www.dgrhoads.com/files1/Estab-Perf-Stds-Dec2004.pdf
59. Ehrmeyer SS, Laessig RH. Has compliance with CLIA requirements really
improved quality in US clinical laboratories? Clin Chim Acta
2004;346(1):37-43.
60. Niwa M. Why not use the standard deviation index as a common scale for data
quantification? Clin Chem 1987;33(7):1294-5.
61. Hanok A, Kuo J. Stability of a reconstituted serum for the assay of fifteen hemical
constituents. Clin Chem 1968;14(1):58-69.
62. Boyanton BL, Jr., Blick KE. Stability studies of twenty-four analytes in human
Plasma and Serum. Clin Chem 2002;48(12):2242-7.
63. Heins M, HeiP W, Withold W. Storage of serum or whole blood samples? Effects
of time and temperature on 22 serum analytes. Eur J Clin Chem Clin
Biochem 1995;33:231-8.
64. Özbek B, Karavelioğlu H. Studies on the stability of various components of
control serums prepared in lyophilized form. [Online]. 2010 [cited 2010
Aug 13]; Available from: URL:http://www.sigma.yildiz.edu.tr/2001-1-4-
tam.DOC
65. Smith AF, Fogg BA. Possible mechanisms for the increase in activity of alkaline
phosphatase lyophilized control material. Clin Chem 1972;18(12):1518-
23.
66. Ziegenhorn J, Neumann U, Hagen A, Bablok W, Strinshoff K. Kinetic enzymatic
method for automated determination of glucose in blood and serum. J Clin
Chem Clin Biochem 1977;15(1):13-9.
67. Mitchell FL. Automation in clinical chemistry: developments and recent trends. J
Automat Chem 1978;1(1):7-9.
APPENDIX A
External Quality Assessment Scheme in Clinical Chemistry for
District Laboratories in Bhutan
(Questionnaire)
Note: this questionnaire is to be filled up by every willing participant after signing the
consent form. The participant's private information will be kept confidential. Fill up and
tick the most appropriate answer and write down few lines of answers wherever required.
Section I: hospital information
1. Name of hospital
2. Category of hospital
i) National Referral Hospital
ii) Regional Referral Hospital
iii) District Hospital
iv) Bed size of the hospital
Section II: laboratory information
1. Name of chemistry analyzer with the name of Supply Company and manufacturing
country.
Name of chemistry
Name of Supply
Manufacturing
Manual/automated
analyzers in use
company
country
or semi-automated?
2. Name of the pipette used for the reconstitution
a. Volumetric glass pipette
b. Adjustable micropipette
c. Others
3. Name of reagents used for routine clinical chemistry tests with the name of Supply
Company and the manufacturing country.
Name of reagents
Supply company
Country of origin
4. Write the methods/principle and volume of sample required per run for the routine
tests in your clinical chemistry.
Routine Tests
Units
Methods/principle
Sample volume
per run
SECTION III: Educational information
1. Total staff in chemistry laboratory
2. Number of Pathologists
3. Number of Medical technologist
4. Number of Laboratory technician
5. Number of stuff trained in quality assurance
SECTION IV: Laboratory Quality Control
A. External Quality assurance activities
1. Is your laboratory member of any EQAS organization? Tick “yes” if you are
member or “no” if not.
1) Yes
2) No
2. When did your laboratory fist participate in EQAS?
Day Month Year
3. What is the name of EQA organization to which your laboratory s a member?
4. What type of QC material your lab is provided by the EQA provider?
1) Liquid form
2) Lyophilized
3) Others
5. How often do you get EQA sample?
1) Weekly
2) Monthly
3) Other (specify)
6. At what temperature do you store your EQAS sample when you receive from the
EQAS provider?
1) Room temp
2) At 4 -8
0
C
3) At -20 -
-
70
0
C
7. How many levels of control material you are provided by your EQA Program?
1. Level I (high)
2. Level II (normal)
3. Level III (low)
8. If your laboratory has not been participated in any EQA program, do you wish to
participate in National EQAS if established?
1) Yes
2) No
If not, why?
B. Internal Quality Control (IQC) activities
1. Is your IQC prepared in your laboratory? If not, name the internal quality
control currently in use in your laboratory with the name of the supply company.
1.
Yes
2.
No
Name of IQC material
Name of supply company/manufacturer
2. How many levels of IQC do you use in chemistry analysis?
(1) Level I (High)
2) Level II (Normal)
3) Level III (Low)
3. How do you check the stability of your reagents and IQC materials?
1) Looking at expiry date
2) Looking at the SD on the Q.C chart
4) Others (specify)
4. Types of IQC material used for chemistry tests.
1) Liquid
2) Lyophilized
3) Others (specify)
Do you calculate the IQC range for your laboratory?
1. Yes
2. No
If not, why?
SECTION V: safety and precaution in the laboratory
1. Do you need to wear gloves in handling EQA and IQC material?
1. Yes
2. No
2. How do you discard your control materials and vials after use?
1) Vials are reused for other purposes after sterilization
2) Discard into dust bin without doing anything.
3) Disinfect first and then discard into a designated place
4) Vials are reused after rinsing with distilled water
I hereby certify that to the best of my knowledge, the information supplied is
accurate, complete and current that I am an in charge of the laboratory and fully
responsible in completing this questionnaire.
Questionnaire completed by
Sign………………………......
Name……………………........
(Laboratory In-charge)
Date........................................
I hereby verify that to the best of my knowledge the information supplied by my
laboratory In-charge is accurate, complete and current and that I am the head of the
hospital that is duly authorized to sign this completed questionnaire.
Questionnaire Verified by:
Sign..........................................
Name……………………….....
(District Medical Officer)
Date..........................................
APPENDIX B
External Quality Assessment Scheme in Clinical Chemistry for District
Laboratories in Bhutan
(Report form)
Please, Fil l up this form with inforamtion provided on the vial and record your results
correctly.
Sample No.
Date of analysis
Date of sample receipt
Sample storage
i) 25-30
o
C
Analytes
Results
Units
iii) 2 – 8
o
C
Glucose
Mg/dl
BUN
Mg/dl
Creatinine
Mg/dl
Pipettes for reconstitution
AST
U/L
i) Micropipette
ALT
U/L
ALP
U/L
ii) Volumetric pipette
Total bilirubin
Mg/dl
iii) Graduated pipette
Total protein
g/dl
Albumin
g/dl
iv) Others (specify)
Color of the reconstituted control material …………………………………….
..………………………….
..………………………….
Date & Signature of analyst
Date & Signature of Verifier
APPENDIX C
EXTERNNAL QULAITY ASSESSMENT SCHEME IN CLINICAL
CHEMISTRYFOR DISTRICT LABORATORIES IN BHUTAN
(Performance Report for the Participants)
Lab Code No Next Dateline
Date of Data Analysis
Date of reporting
In compliance to the ethics of the research, the identity of a participant is kept confidential
Trial No.
Your
Overall
Overall
xxxx
Unit
Your Scores
Results
Mean
SD
Analytes
VIS
Interpretation
Glucose
mg/dl
BUN
mg/dl
Creatinine
mg/dl
AST
U/L
ALT
U/L
ALP
U/L
T. Bilirubin
mg/dl
T. Protein
g/dl
Albumin
g/dl
Key words: NR = No reagent; SD =Standard Deviation; VIS=Variance Index Score;
CPK=Capability Index , BUN=Blood Urea Nitrogen; AST =Aspartate
Aminotransferase; ALT =Alanine Aminotransferase, ALP =Alkaline Phosphatase;
Interpretation of VIS scores
0-50 = Excellent; 51-100 = Very good; 101-150 = good; 151-200 = Acceptable; 201-
250 =Needs improvement; VIS >250: Unacceptable, immediate action should be
taken to correct the errors.
Your results in respect to other laboratories: The graphs are plotted with analyte
values on the x-axis versus no. of laboratories on the y-axis to show the distribution of
the results for all the laboratories. The position of your results of each analyte is
shown by an arrow. The figure on top of each bar is the number of laboratories
attaining the results with the range given at its bottom.
Reported by : Report verified by :
Rixin Jamtsho Assist. Prof. Dr. Wilairat Nuchpramool
APPENDIX D
EXTERNAL QUALITY ASSESSMENT SCHEME IN CLINICAL
CHEMISTRY FOR DISTRICT LABORATORIES IN BHUTAN (EQAC-
DLB) (Package Insert)
Name of Hospital: ……………………
Trial No. 001-024 Mfg: 11/11/09
1. System information
For use on Screen master 3000, BA-88, Beckam Coulter and Hitachi 911 & 912
2. Frequency of analysis: Every month for one year as given in the table below.
3. Summary
Control serum is prepared by lyophilizing a pool of fresh-frozen human serum to provide
constituents in a stable form. They should be analyzed in similar manner as patient‟s
serum.
4. Schedule for sample measurement
Two samples should be analyzed every month within the first week of scheduled
month (except the 1
st
paired sample which has to be analyzed at the end of November
due to delay in sending the control material. The report should be sent to the
coordinator in the same week. Results should be recorded with all the necessary
information (code and trial number). The dateline for the analysis and reporting has
been given on the each paired sample for the convenience of the sample handlers.
Follow the table for monthly analysis of the sample pairs as follows:
Table shows the sequence of the sample trial numbers for monthly analysis
Month
Nov
Dec
Jan
Feb
Mar
April
May
June
July
Aug
Sept
Oct
Year
2009
2009
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
001
003
005
007
009
011
013
015
017
019
021
023
T. No.
002
004
006
008
010
012
014
016
018
020
022
024
5. Safety in handling control material
All human material should be considered potentially infectious. These materials were
tested and shown to be free from HBsAg and antibodies to HIV. However, this
doesn‟t rule out the risk of infection with absolute certainty. Therefore, the necessary
safety precautions should not be neglected. Laboratory personnel handling should
keep in mind some important safety measures as follows:
1) This product, like all materials of human origin, should be handled as
potentially infectious biological material. The product is for in vitro diagnostic
use only
2) Do not pipette by mouth
3) Wear Personal Protective Equipment (PPE) and avoid physical contact with
material.
4) Wash hands thoroughly after handling.
5) Absorb spill with inert material/sorbent and decontaminate the spill site
following standard procedures for hazardous spills.
5. Reconstitution technique
Take out the lyophilized sample and stand for 30 minutes at room temperature.
Carefully open one vial avoiding the loss of lyophilisate
Carefully pipette in exactly 2.0 ml of deionized water at 20-25
o
C with volumetric pipette
Close the bottle and dissolve the contents completely by occasional gentle swirling within 30
minutes. Do not shake the vial.
6. Precautions
Only volumetric glass pipette should be used to dispense distilled water for reconstitution
Avoid air bubbles in the pipette while dispensing distilled water
After reconstitution, allow the control material to stand for 1hr at 20-25
o
C to reactivate
alkaline phosphatase
Avoid exposure to direct sun light
Use only de-ionized or double distilled water to avoid contamination
Invert the vial after 30 minutes mix thoroughly to make sure that no lyophilisate is adhered to
the wall of the bottle.
7. Storage
Lyophilized serum should be stored at 2-8
o
C and freezing -20
o
C can prolong the
shelf life.
7. Requirements
Graduated glass pipette (1-10ml)
Chemistry analyzer
Refrigerator (2-8
o
C)
Double distilled water
Disinfectant solutions, Gloves and gowns
********************************************************************
BIOGRAPHY
NAME Mr. Rixin Jamtsho
DATE OF BIRTH 3
rd
November, 1976
PLACE OF BIRTH Mongar, Bhutan
INSTITUTE ATTENDED Rajiv Gandhi University, Bangalore, India
2000-2003, Bachelor of
Science in Medical Technology
Mahidol University 2008-2010
Master of Science
(Medical Technology)
SCHOLARSHIP Royal Government of Bhutan
POSITION AND OFFICE Jigme Dorji Wangchuk National
Referral Hospital, Thimphu, Bhutan.
Position: Medical Technologist
HOME ADDRESS Changzamtok, Thimphu
Tel. +975177606984
E-mail: rixinjams@hotmail.com
