Clin Chem Lab Med 2017; 55(3): 348–357

IFCC Paper

Laura Sciacovelli, Giuseppe Lippi, Zorica Sumarac, Jamie West, Isabel Garcia del Pino Castro,
Keila Furtado Vieira, Agnes Ivanov and Mario Plebani*, on behalf of the Working Group
“Laboratory Errors and Patient Safety” of International Federation of Clinical Chemistry
and Laboratory Medicine (IFCC)

Quality Indicators in Laboratory Medicine: the

status of the progress of IFCC Working Group
“Laboratory Errors and Patient Safety” project
DOI 10.1515/cclm-2016-0929 and the population they serve. Moreover, it also depends
Received October 13, 2016; accepted November 1, 2016; previously on the choice of the events to keep under control and
published online December 19, 2016 the individual procedure for measurement. Although
many laboratory professionals believe that the systemic
Abstract: The knowledge of error rates is essential in
use of QIs in Laboratory Medicine may be effective in
all clinical laboratories as it enables them to accurately
decreasing errors occurring throughout the total test-
identify their risk level, and compare it with those of
ing process (TTP), to improve patient safety as well as to
other laboratories in order to evaluate their performance
satisfy the requirements of International Standard ISO
in relation to the State-of-the-Art (i.e. benchmarking)
15189, they find it difficult to maintain standardized and
and define priorities for improvement actions. Although
systematic data collection, and to promote continued
no activity is risk free, it is widely accepted that the risk of
high level of interest, commitment and dedication in
error is minimized by the use of Quality Indicators (QIs)
the entire staff. Although many laboratories worldwide
managed as a part of laboratory improvement strategy
express a willingness to participate to the Model of QIs
and proven to be suitable monitoring and improvement
(MQI) project of IFCC Working Group “Laboratory Errors
tools. The purpose of QIs is to keep the error risk at a
and Patient Safety”, few systematically enter/record
level that minimizes the likelihood of patients. However,
their own results and/or use a number of QIs designed
identifying a suitable State-of-the-Art is challenging,
to cover all phases of the TTP. Many laboratories justify
because it calls for the knowledge of error rates meas-
their inadequate participation in data collection of QIs
ured in a variety of laboratories throughout world that
by claiming that the number of QIs included in the MQI
differ in their organization and management, context,
is excessive. However, an analysis of results suggests
that QIs need to be split into further measurements. As
*Corresponding author: Mario Plebani, Department of Laboratory the International Standard on Laboratory Accreditation
Medicine, University Hospital, Padova, Italy, and approved guidelines do not specify the appropriate
E-mail: [email protected]. number of QIs to be used in the laboratory, and the MQI
http://orcid.org/0000-0002-0270-1711
project does not compel laboratories to use all the QIs
Laura Sciacovelli: Department of Laboratory Medicine, University
Hospital, Padova, Italy
proposed, it appears appropriate to include in the MQI
Giuseppe Lippi: Laboratory of Clinical Chemistry and Hematology, all the indicators of apparent utility in monitoring criti-
University Hospital, Verona, Italy cal activities. The individual laboratory should also be
Zorica Sumarac: Centre for Medical Biochemistry, Clinical Centre of able to decide how many and which QIs can be adopted.
Serbia, Belgrade, Serbia In conclusion, the MQI project is proving to be an impor-
Jamie West: Department of Clinical Biochemistry and Immunology,
tant tool that, besides providing the TTP error rate and
Peterborough and Stamford Hospitals NHS Foundation Trust,
Peterborough, UK spreading the importance of the use of QIs in enhancing
Isabel Garcia del Pino Castro: Area Laboratory, Complexo patient safety, highlights critical aspects compromising
Hospitalario, Universitario A Coruña, A Coruña, Spain the widespread and appropriate use of QIs.
Keila Furtado Vieira: Pontificia Universidade Católica de São Paulo,
Sorocaba, São Paulo, Brazil Keywords: extra-analytical phases; laboratory error;
Agnes Ivanov: Tartu University Hospital, Tartu, Estonia quality indicators; total testing process.
 Sciacovelli et al.: Quality Indicators in Laboratory Medicine      349

Introduction systematic data collection along with a high level of

interest, commitment and dedication from the entire
The growing awareness of the importance of the extra- staff. In order to overcome these problems and identify
analytical phases in generating reliable laboratory infor- the State-of-the-Art concerning errors occurring in the
mation has prompted clinical laboratories to closely TTP, the Working Group “Laboratory Errors and Patient
observe all activities in the total testing process (TTP) in Safety” (WG-LEPS) of the International Federation of
order to identify all possible error risks. According to the Clinical Chemistry and Laboratory Medicine (IFCC) has,
ISO Guide 73 : 2009 “Risk is often expressed in terms of a since 2008, implemented a project aimed at defining a
combination of the consequences of an event (including common Model of QIs (MQI), a harmonized method for
changes in circumstances) and the associated likelihood data collection, managed as an External Quality Assur-
of occurrence” [1]. Frequently occurring events of low level ance Program (EQAP) in which confidentiality is guaran-
severity are associated with higher risks, but high severity teed [9–11].
events, even if isolated/rare can incur even higher risks An MQI issued by a Consensus Conference held in
[2]. It is therefore mandatory for the identification and Padova in 2013 and used since 2014. involves a priority
monitoring of activities with a high risk of error to become score being assigned to each indicator, or assisting labo-
a modus operandi in all laboratory procedures. According ratories to gradually introduce QIs into routine practice.
to the ISO 15189:2012, clinical laboratories should identify A criterion to identify Quality Specifications (QSs) for
critical TTP activities and implement Quality Indicators assessing laboratory performance has also been proposed
(QIs) in order to highlight and monitor errors when they [12, 13].
occur. QIs, managed as a part of laboratory improvement
strategy have proven to be a suitable tool in monitoring
and achieving improvement [3], their ultimate purpose
being to keep the error risk at a level that minimizes the
Aim
likelihood of patient harm, given that no activity is com-
This article describes the state of progress of the MQI
pletely risk-free. Data available in the literature demon-
project. The results reported for each indicator are as
strate that the effectiveness of this tool is closely linked to
follows:
the list of QIs chosen, and to: a) data collection method,
–– statistical data of QIs data collected in the 2014, 2015
b) data processing procedure in use, c) appropriate analy-
and 2016 (6 months);
sis of results, and d) an understanding of the priorities for
–– statistical data of sigma values for data collected in
corrective actions according to performance of the various
the years 2014, 2015 and 2016 (first 6 months);
QIs [4–7].
–– criteria used to define QSs.
The knowledge of error rates is essential for any
clinical laboratory as it enables the service to correctly
The critical aspects highlighted by participants and
identify of its own risk level, and to compare it with
emerging during the use of QIs use are also described, and
those of other laboratories in order to evaluate its perfor-
future trends considered.
mance in relation to the State-of the-Art (i.e. benchmark-
ing) and identify the priorities for improvement actions.
Nevertheless, identifying a suitable State-of-the-Art is
a challenging issue, calling for the knowledge of error Methodology
rates measured worldwide in laboratories that have dif-
ferent organizational and management aspects and con- All laboratory data collected in 2014, 2015 and in the first
texts, and serve different populations. Moreover, it also 6 months of 2016, were processed on a yearly basis, and the
depends on the choice of events to keep under control values of the 25th, 50th and 75th percentiles calculated.
and the procedure that an individual laboratory uses for A similar procedure was used to estimate sigma
measurement. values, the short-term formula being applied for the QIs
Although many laboratory professionals believe that expressed in percentages [14].
the systemic use of QIs in Laboratory Medicine may be The criterion used to define QSs is based on percentile
effective in decreasing errors occurring throughout TTP values estimated according to laboratory results. Three
with a view to enhancing patient safety and meeting levels of performance were identified:
the requirements of International Standard ISO 15189 –– high, 25th percentile value, representing the best
[8], they find it difficult to maintain standardized and performance;
350      Sciacovelli et al.: Quality Indicators in Laboratory Medicine

–– medium, 50th percentile value, representing the to improve over time, the performance goals need to be
more frequent/common performance; regularly reviewed (e.g. annually) by analysing the error
–– low, 75th percentile representing the worst rate recorded. The knowledge of QSs enables clinical labo-
performance. ratories to identify the most suitable corrective/improve-
ment actions and the relative priorities, whereas it may be
When the QIs results measured the desirable events excessively challenging to focus improvement projects on
(­Post-Comm, Supp-Train, Supp-Cred, Supp-Phys, Supp- all the activities being monitored.
Pat), the high level of performance corresponded to the
75th percentile and the low, to the 25th percentile. When
the percentile values coincided, it was possible to use a
single value. Discussion
Although many laboratories worldwide expressed their
Results willingness to participate in the MQI project, only a few
of them systematically entered their own results or used
Table 1 shows all the findings for QIs (in particular, 25th, a number of QIs designed to cover all phases of the TTP.
50th and 75th percentiles), obtained from data collected The main QIs used, classified according to the phase
by laboratories that consistently participate in the MQI of the TTP, are as follows:
project. Overall, data were received from 59 laboratories –– pre-analytical phase: a) unsuitable samples (haemo-
in: Argentina, 2; Austria, 1; Brazil, 1; Estonia, 2; Germany, lysed, clotted, inappropriate sample-anticoagulant
1; Great Britain, 2; India, 2; Italy, 16; Republic of China, 2; volume ratio, insufficient volume, wrong container,
Republic of Croatia, 6; Spain, 2; Switzerland, 2; Serbia, 19; unlabelled, inappropriate type, not received) and b)
Uruguay, 1. misidentified errors (requests and samples);
The short-term sigma for QIs, expressed as a percent- –– intra-analytical phase: a) unacceptable performance
age, was estimated in order to identify the quality level of in EQAs-PT and b) tests with inappropriate internal
the processes monitored (Table 1). The sigma quality level quality control (IQC) performance;
provides information on the frequency of the occurrence/ –– post-analytical phase: a) incorrect reports issued and
risk of the various defects. A higher sigma quality level b) inappropriate TAT (reports delivered outside the
indicates that a process is less likely to generate problems, specified time, critical values notified after a consen-
thus also indicating that the need for checking and inspec- sually agreed time, Potassium TAT).
tion is reduced, costs are lower, and customer satisfaction
enhanced. The estimation of sigma value is not applicable For QIs of Outcome Measures and Support Processes, all
to QIs results that cannot be expressed in percentages (i.e. indicators proposed in MQI appear to be used in a similar
minutes or numbers). fashion, but only by a small number of laboratories.
The criterion adopted to identify the QSs for each indi- Many laboratories justify their inadequate participa-
cator includes the definition of three different performance tion in QI data collection by citing the ‘excessive’ number
goals (low, medium, high) according to laboratory results, of QIs included in the MQI. However, it is important to
thus highlighting the most recent error rates collected bear in mind that:
at a particular time. Information on a performance level –– QIs should monitor all critical aspects of the TTP, as
based on measures allows each laboratory to establish stated by ISO 15189:2012, and therefore several QIs
and compare its placing with that of other laboratories, need to be decided upon;
thus making it possible to plan improvement actions. The –– to obviate any confusion between indicator and meas-
use of the 75th percentile as the lower limit seems to be a urements, different measures are often required to
more practical approach, indicating that performance was ensure that an indicator is appropriately monitored.
poor in less than 25% of participating laboratories. In fact, It is advisable to split an indicator into different
a high percentage of unsatisfactory performances may measures in order to consider all the events causing
discourage some laboratories from attempting to improve a specific error, and to benchmark data entered by
quality. On the other hand, if laboratories see they have different laboratories. Some laboratories, for exam-
achieved a higher goal, they are not motivated to under- ple, offer their service to outpatients and inpatients,
take improvement actions. Since the improvement actions while others perform analyses for only one of these
implemented by the different laboratories are expected two patient groups. Any comparison of data entered
 Sciacovelli et al.: Quality Indicators in Laboratory Medicine      351

Table 1: Model of quality indicators: results from 2014 to 2016 (6 months).

Indicator   Year  Results  Note

 
Percentile  Sigma
 
 
Priority – measure 25th  50th  75th 25th  50th  75th

Key-processes: Pre-Anaytical phase                

 Misidentification errors                
1   – Percentage of: Number of misidentified requests/   2014  0.005  0.0345  0.2857  4.2  4.6  5.0 
Total number of: requests. (Pre-MisR)   2015  0  0.016  0.154  4.3  4.7  5.1 
  2016  0.0015  0.0365  0.1595  4.4  4.7  5.0 
1   – Percentage of: Number of misidentified samples/   2014  0  0.013  0.039  4.7  4.9  5.1 
Total number of samples. (Pre-MisS)   2015  0.001  0.0195  0.063  4.7  4.9  5.1 
  2016  0  0.031  0.056  4.5  4.8  4.9 
1   – Percentage of: Number of samples with fewer than 2   2014  0.0012  0.06  0.294  4.1  4.5  4.9 
identifiers initially supplied/Total number of samples.   2015  0  0.01  0.1685  4.1  4.4  4.7 
(Pre-Iden)   2016  0  0.0985  0.2825  3.0  4.4  4.6 
1   – Percentage of: Number of unlabelled samples/Total   2014  0  0.01  0.0355  4.7  4.9  5.2 
number of samples. (Pre-UnlS)   2015  0  0.007  0.0252  4.7  5.0  5.2 
  2016  0  0.03  0.012  4.7  5.2  5.2 

Test transcription errors                

1   – Percentage of: Number of outpatients requests with   2014  0  0.118  0.654  3.8  4.1  4.5 
erroneous data entry (test name)/Total number of   2015  0  0.183  0.5267  4.0  4.2  4.4 
outpatients requests. (Pre-OutpTN)   2016  0  0.132  0.5482  3.8  4.1  4.4 
1   – Percentage of: Number of outpatients requests with   2014  0.0175  0.2995  0.8912  3.8  4.0  4.4 
erroneous data entry (missed test)/Total number of   2015  0  0.2515  0.76  3.8  4.0  4.2 
outpatients requests. (Pre-OutpMT)   2016  0  0.118  0.693  3.8  4.1  4.3 
1   – Percentage of: Number of outpatients requests with   2014  0  0.044  0.3375  4.0  4.3  4.6 
erroneous data entry (added test)/Total number of   2015  0  0  0.1132  4.3  4.5  4.8 
outpatients requests. (Pre-OutpAT)   2016  0  0  0.0935  4.6  4.6  4.7 
1   – Percentage of: Number of inpatients requests with   2014  0  0.07  0.567  3.9  4.2  4.6 
erroneous data entry (test name)/Total number of   2015  0  0  0.135  4.1  4.4  4.7 
inpatients requests. (Pre-InpTN)   2016  0  0  0.066  4.4  4.6  4.9 
1   – Percentage of: Number of inpatients requests with   2014  0  0.1205  0.504  3.9  4.2  4.6 
erroneous data entry (missed test)/Total number of   2015  0  0.013  0.1055  4.2  4.6  4.8 
inpatients requests. (Pre-InpMT)   2016  0  0.012  0.114  3.7  4.6  4.6 
1   – Percentage of: Number of inpatients requests with   2014  0  0.224  0.671  3.8  4.1  4.4 
erroneous data entry (added test)/Total number of   2015  0  0.013  0.681  3.9  4.2  5.0 
inpatients requests. (Pre-InpAT)   2016  0  0.0305  0.9335  3.8  3.9  4.6 

Incorrect sample type                

1   – Percentage of: Number of samples of wrong or   2014  0  0.004  0.027  4.8  4.9  5.2 
inappropriate type (i.e. whole blood instead of   2015  0  0.002  0.034  4.7  4.9  5.2 
plasma)/Total number of samples. (Pre-WroTy)   2016  0  0.002  0.02  4.6  4.9  5.2 
1   – Percentage of: Number of samples collected in wrong   2014  0.002  0.013  0.0327  4.8  5.0  5.2 
container/Total number of samples. (Pre-WroCo)   2015  0.004  0.012  0.029  4.9  5.0  5.2 
  2016  0.004  0.014  0.0295  4.9  4.9  5.2 

Incorrect fill level                

1   – Percentage of: Number of samples with insufficient   2014  0.012  0.032  0.0885  4.6  4.8  5.0 
sample volume/Total number of samples. (Pre-InsV)   2015  0.012  0.027  0.07  4.6  4.9  5.0 
  2016  0.018  0.041  0.109  4.5  4.7  5.0 
1   – Percentage of: Number of samples with inappropriate   2014  0.064  0.267  0.589  4.0  4.2  4.6 
sample-anticoagulant volume ratio/Total number of   2015  0.1192  0.342  0.6047  4.0  4.2  4.5 
samples with anticoagulant. (Pre-SaAnt)   2016  0.0845  0.2365  0.5885  4.0  4.3  4.6 

Unsuitable samples for transportation and storage problems

1   – Percentage of: Number of samples not received/Total   2014  0.06  0.261  1.123  3.8  4.3  4.6 
number of samples. (Pre-NotRec)   2015  0.0875  0.493  1.089  3.8  4.0  4.6 
  2016  0.2175  0.6995  1.020  3.8  3.9  4.3 
352      Sciacovelli et al.: Quality Indicators in Laboratory Medicine

Table 1 (continued)

Indicator   Year  Results  Note

 
Percentile  Sigma
 
 
Priority – measure 25th  50th  75th 25th  50th  75th

1   – Percentage of: Number of samples not properly stored   2014  0  0  0.027  4.8  4.9  5.0 
before analysis/Total number of samples. (Pre-NotSt)   2015  0  0  0.008  4.9  5.0  5.4 
  2016  0  0  0.009  4.9  5.1  5.2 
1   – Percentage of: Number of samples damaged during   2014  0  0  0.002  4.9  5.2  5.2 
transportation/Total number of samples. (Pre-DamS)   2015  0  0  0.003  5.2  5.2  5.5 
  2016  0  0  0.001  5.2  5.4  5.5 
1   – Percentage of: Number of samples transported at   2014  0  0.002  0.431  3.7  4.1  4.9 
inappropriate temperature/Total number of samples.   2015  0  0.001  0.5305  3.6  3.9  5.2 
(Pre-InTem)   2016  0  0.002  0.584  3.7  3.9  5.3 
1   – Percentage of: Number of samples with excessive   2014  0  0.018  0.564  3.7  4.1  4.9 
transportation time/Total number of samples.   2015  0  0.001  0.181  4.0  4.4  4.9 
(Pre-ExcTim)   2016  0  0.002  0.129  3.9  4.4  4.7 
Contaminated samples                
1   – Percentage of: Number of contaminated samples   2014  0.048  0.2275  1.897  3.4  3.8  4.5 
rejected/Total number of microbiological samples.   2015  0.163  1.481  3.847  3.3  3.6  4.2 
(Pre-MicCon)   2016  0.1457  1.095  5.405  3.1  3.7  4.4 

Sample haemolysed                
1   – Percentage of: Number of samples with free   2014  0.437  0.866  1.548  3.7  3.9  4.1 
Hb>0.5 g/L (clinical chemistry)/Total number of   2015  0.492  1.059  1.854  3.6  3.8  4.1 
samples (clinical chemistry) · (Pre-Hem)   2016  0.555  1.405  2.567  3.4  3.7  4.0 

Samples clotted                
1   – Percentage of: Number of samples clotted/Total   2014  0.11  0.317  0.611  4.0  4.2  4.5 
number of samples with an anticoagulant. (Pre-Clot)   2015  0.165  0.98  0.5205  4.1  4.2  4.4 
    2016  0.108  0.299  0.459  4.1  4.2  4.6 

Inappropriate test requests                

2   – Percentage of: Number of requests without clinical   2014  0.750  7.436  59.03  1.0  2.7  3.4 
question (outpatients)/Total number of requests   2015  1.183  2.598  18.06  2.3  3.3  3.7 
(outpatients). (Pre-Quest)   2016              Not available due
to poor results

Inappropriate time in sample collection                

2   – Percentage of: Number of samples collected at   2014  0  0.075  0.0432  4.6  4.9  5.2 
inappropriate time of sample collection/Total number   2015  0  0  0.346  4.0  4.1  4.2 
of samples. (Pre-InTime)   2016              Not available due
to poor results

Unintelligible requests                
3   – Percentage of: Number of unintelligible outpatients   2014  0  0.363  1.137  3.6  3.8  4.2 
requests/Total number of outpatients requests.   2015  0  0  0.47  3.7  4.0  4.2 
(Pre-OutUn)   2016  0  0  0.104  3.9  4.3  4.6 
3   – Percentage of: Number of unintelligible inpatients   2014  0  0.069  0.406  4.0  4.2  4.4 
requests/Total number of inpatients requests.   2015  0  0  0.012  4.0  4.3  4.5 
(Pre-InpUn)   2016              Not available due
to poor results

Inappropriate requests                
4   – Percentage of: Number of inappropriate requests, with   2014  0.0457  0.757  2.163  3.5  3.6  4.3 
respect to clinical question (outpatients)/Number of   2015  1.489  1.601  2.93  3.4  3.6  3.7 
requests reporting clinical question (outpatients).   2016              Not available due
(Pre-OutReq) to poor results
4   – Percentage of: Number of inappropriate requests, with   2014  0  0.292  4.79  2.4  3.4  4.0 
respect to clinical question (inpatients)/Number of   2015  0  1.842  5.457  2.8  3.1  3.2 
requests reporting clinical question (inpatients).   2016              Not available due
(Pre-InReq) to poor results
 Sciacovelli et al.: Quality Indicators in Laboratory Medicine      353

Table 1 (continued)

Indicator   Year  Results  Note

 
Percentile  Sigma
 
 
Priority – measure 25th  50th  75th 25th  50th  75th

Key-processes: Intra-Anaytical phase                

Test with inappropriate ICQ performances                
1   – Percentage of: Number of tests with CV% higher   2014  0  0.005  15.71  2.1  2.5  2.7 
than selected target, per year/Total number of tests   2015  0  2.26  12.5  2.7  2.7  3.4 
with CV% known for at least: Glucose; Creatinine;
Potassium; C-Reactive Protein; Troponin I or Troponin
T; TSH; CEA; PT (INR); Haemoglobin (HB). (Intra-Var)

Tests not covered by an EQA-PT control                

1   – Percentage of: Number of tests without EQA-PT   2014  14.82  31.82  47.31  1.6  2.0  2.5 
control/Total number of tests in the menu.   2015  15.28  24.93  34.4  1.9  2.2  2.5 
(Intra-EQA)

Unacceptable performances in EQA-PT schemes                

1   – Percentage of: Number of unacceptable performances   2014  0.769  2.541  4.615  3.0  3.3  3.5 
in EQAS-PT Schemes, per year/Total number of   2015  1.89  2.4  3.134  3.3  3.4  3.6 
performances in EQA Schemes, per year. (Intra-Unac)
3   – Percentage of: Number of unacceptable performances   2014  0  0  10.36  2.0  2.3  2.6 
in EQAS-PT Schemes per year occurring to previously   2015  0  0  3.17  3.0  3.1  3.2 
treated cause/Total number of unacceptable
performances. (Intra-PPP)

Data transcription errors                

1   – Percentage of: Number of incorrect results for   2014  0  0  0.036  4.6  4.9  5.0 
erroneous manual transcription/Total number of   2015  0  0  0.003  4.5  5.1  5.2 
results requiring manual transcription. (Intra-ErrTran)   2016  0  0  0  5.2  5.4  5.5 
1   – Percentage of: Number of incorrect results for   2014  0  0  0  5.0  5.0  5.0 
information system problems-failures/Total number   2015  0  0  0  4.9  4.9  4.9 
of results. (Intra-FailLIS)   2016  0  0  0  5.2  5.2  5.2 

Key-processes: Post-Anaytical phase                

Inappropriate turnaround times                
1   – Percentage of: Number of reports delivered outside the   2014  0  0.035  0.554  3.6  4.3  4.7 
specified time/Total number of reports. (Post-OutTime)   2015  0  0.224  1.95  3.3  4.2  4.4 
  2016  0  0.21  1.79  2.8  3.9  4.4 
1   – Turn Around Time (minutes) of Potassium (K) at 90th   2014  48  49.6  60        Estimate of
percentile (STAT). (Post-PotTAT) sigma value not
applicable
  2015  56.5  73  89.34       
  2016              Not available due
to poor results
1   – Turn Around Time (minutes) of International   2014  42  45  49.5       
Normalized Ratio (INR) value at 90th percentile (STAT).   2015  46  48.97  59.5       
(Post-INRTAT)   2016              Not available due
to poor results
1   – Turn Around Time (minutes) of White Blood Cell Count   2014  17.5  23  26        Not available due
(WBC) at 90th percentile (STAT). (Post-WBCTAT) to poor results
  2015              Not available due
to poor results
  2016              Not available due
to poor results
1   – Turn Around Time (minutes) of Troponin I (TnI) or   2014  51  53  71.5       
Troponin T (TnT) at 90th percentile (STAT). (Post-TnTAT)   2015  47.5  51  62.93       
354      Sciacovelli et al.: Quality Indicators in Laboratory Medicine

Table 1 (continued)

Indicator   Year  Results  Note

 
Percentile  Sigma
 
 
Priority – measure 25th  50th  75th 25th  50th  75th

Incorrect laboratory reports                

1   – Percentage of: Number of incorrect reports issued by   2014  0  0  0.041  4.7  4.8  4.9 
the laboratory/Total number of reports issued by the   2015  0  0.01  0.03  4.8  4.9  5.0 
laboratory. (Post-IncRep)   2016  0  0.006  0.017  4.9  5.0  5.2 

Notification of critical values                

  – Percentage of: Number of critical values of inpatients   2014  0  1.12  8.333  2.1  3.0  3.4 
notified after a consensually agreed time (from result   2015  0  0.765  6.989  1.8  3.1  3.5 
validation to result communication to the clinician)/   2016              Not available due
Total number of critical values of inpatients to to poor results
communicate.
(Post-InpCV)
1   – Percentage of: Number of critical values of outpatients   2014  0  0  22.596  1.3  2.2  2.7 
notified after a consensually agreed time (from result   2015              Not available due
validation to result communication to the clinician)/ to poor results
Total number of critical values of outpatients to   2016              Not available due
communicate. (Post-OutCV) to poor results

Interpretative comments                
4   – Percentage of: Number of reports with interpretative   2014  0.156  34.19  60.625  1.7  1.9  3.9  Best performance:
comments, provided in medical report, impacting 75th percentile
positively on patient's outcome/Total number of   2015              Not available due
reports with interpretative comments. (Post-Comm) to poor results
  2016              Not available due
to poor results

Results notification (TAT)                

4   – Time (from result validation to result communication                 Not available due
to the clinician) to communicate critical values of to poor results
inpatients (minutes). (Post-InCVT)
4   – Time (from result validation to result communication                 Not available due
to the clinician) to communicate critical values of to poor results
outpatient (minutes). (Post-OutCVT)                

Outcome measures                
Sample recollection                
1   – Percentage of: Number of outpatients with recollected   2014  0  0  0.0495  4.4  4.7  4.9 
samples for laboratory errors/Total number of   2015  0  0.046  0.399  4.1  4.7  4.9 
outpatients. (Out-RecOutp)   2016  0  0.046  0.369  4.1  4.3  4.8 
1   – Percentage of: Number of inpatients with recollected   2014  0  0  0  4.5  4.6  4.9 
samples for laboratory errors/Total number of   2015  0  0  0.038  4.2  4.7  4.9 
inpatients. (Out-RecInp)   2016  0  0  0.106  4.2  4.5  4.7 

Inaccurate results                
1   – Percentage of: Number of inaccurate results released/   2014  0  0  0  4.5  4.9  5.0 
Total number of results released. (Out-InacR)   2015  0  0  0  5.0  5.0  5.0 
  2016  0  0  0  5.0  5.0  5.0 
               
Support processes                
Employee competence                
2   – Number of training events organized for all staff,                
per year (Supp-Train) 
2   – Percentage of: Number of credits obtained by   2014  88.08  100  100        Best performance:
employee, per year/Total number of credits to be 75th percentile
obtained, per year. (Supp-Cred)   2015  31.46  64.06  94.231        Best performance:
75th percentile
 Sciacovelli et al.: Quality Indicators in Laboratory Medicine      355

Table 1 (continued)

Indicator   Year  Results  Note

 
Percentile  Sigma
 
 
Priority – measure 25th  50th  75th 25th  50th  75th

Client relationships                
2   – Percentage of: Sum of point given in the enquiry to   2014  80  90  96        Better
the question of global satisfaction of the physician/ performance:
Multiplication of the maximum point defined in the 75th percentile
enquiries by the number of enquiries. (Supp-Phys)   2015              Not available due
to poor results
2   – Percentage of: Sum of point given in the enquiry to   2014  80  90  98        Best performance:
the question of global satisfaction of the patient/ 75th percentile
Multiplication of the maximum point defined in the   2015              Not available due
enquiries by the number of enquiries. (Supp-Pat) to poor results

Efficiency of Laboratory Information System                

2   – Number of Laboratory Information System downtime   2014              Not available due
episodes, per year. (Supp-FailLIS) to poor results

by laboratories that process samples from different MQI project does not oblige laboratories to use all QIs pro-
patient populations might generate misleading con- posed, it seems appropriate to include in the MQI all the
clusions. This applies in particular to some indicators indicators that appear useful in monitoring critical activi-
showing wrong procedures performed by different ties. The individual laboratory should be able to decide
personnel as the leading cause of the error. Sample how many, and which, QIs are to be adopted.
collection is a paradigmatic example, wherein the Another aspect biasing the participation in the
error is typically attributable to clinical ward staff for project is related to difficulties in data collection, espe-
inpatient samples and to the laboratory and periph- cially when automated collection is unavailable. The
eral drawing centre personnel for specimens collected laboratory staff may be discouraged and dissatisfied
from outpatients. from manual collection of data, since this activity takes
time and dedication. The design of dedicated software
Rather than deleting some QIs, it might be preferable to for automated data collection could hence stimulate a
revise MQI in order to identify the QIs that should be major involvement of the laboratory staff in the project
split into further measurements. For example, in the [15].
case of haemolysed samples, the error rate of 1.06 esti- However, several real difficulties have been
mated in 2015 included laboratories that used to detect acknowledged in the collection of data with some post-
haemolysis by means of serum indices, visual inspec- analytical QIs. In many cases, identification errors call
tion or other unspecified procedures. An error rate of for enquiry and the active involvement of clinicians/
1.18 was calculated for laboratories using serum indices, nurses, a challenging requirement in from the view-
but only 0.63 for other facilities using visual inspection. point of time and frequency. In other circumstances, it
This clearly indicates that it may be advisable to split seems necessary to better specify which events need to
this indicator into two different measures to prevent measured and how this can be done (i.e. Post-Comm,
misleading conclusions concerning the real burden of interpretative comments impacting positively on
haemolysis. The lesson learnt with this QI implies that it patient’s outcome).
might be better to differentiate the various QSs accord- The laboratories also experienced difficulty in
ing to the specific detection procedure used in the dif- meeting the deadline for collecting and entering data in
ferent laboratories. A similar consideration can be made the MQI-dedicated website. Laboratories are more inclined
for indicators used for tests with a CV% higher than the undertake the retrospective collection of data, with trans-
set target (Intra-Var). mission delayed by months or, in extreme cases, a year. As
As the International Standard on Laboratory Accredi- shown in Table 1, the results of some indicators obtained
tation and approved guidelines do not specify the appro- in 2016  have been excluded due to the low number of
priate number of QIs to be used in the laboratory, and the responses. The failure to comply with these deadlines, in
356      Sciacovelli et al.: Quality Indicators in Laboratory Medicine

turn, further delays the provision of reports to the clinical should be part of a coherent and coordinated quality
laboratories participating the MQI project. improvement strategy, should be constantly reviewed
As regards the sigma values estimate for the QIs of and updated, comply with accreditation requirements
intra-analytical phase, significant improvements have and scientific recommendations, support efforts to con-
been achieved in the last few decades, whereas fewer tinuously improving laboratory performances, enhance
improvements have been made to the extra-analytical the value of both TTP and clinical practice, and be effec-
phases (Table 1). The accurate interpretation of the sig- tive in evaluating patient’s outcome. Additional efforts
nificance of intra-analytical QIs is of crucial importance, should be made to ensure the effective the use of QIs in
as these QIs are not intended to monitor the performance clinical laboratories. The MQI project is proving to be an
of the analytical procedures, but to reflect the manage- important tool that not only provides the TTP error rate
ment of unsatisfactory analytical performances. This and divulges awareness of the value of QIs in enhanc-
observation highlights the need for a greater focus on ing patient safety, but also highlights the more critical
this issue, which is often overlooked. Some laboratories aspects interfering with the widespread and appropri-
manage (or correct) an error at the same time as its occur- ate use of QIs themselves. The dedicated website (www.
rence (i.e. unsatisfactory performance in EQA or IQC), ifcc-mqi.com), already useful for sharing the list of QIs,
thus overriding the underlying cause(s) or disregarding showing the frequency of data collection, and providing
appropriate improvement actions (risk management). valuable information, could be further improved as it is
Due to the type of results and to the lower number of of promise as a tool for connecting participating labora-
responses, a significant sigma value could not be calcu- tories and stakeholders.
lated for the support processes.
In order reduce the error rates in critical TTP proce- Author contributions: All the authors have accepted
dures, the following some initiatives must be undertaken: responsibility for the entire content of this submitted
1) involving Scientific Societies of different countries manuscript and approved submission.
to promote participation of laboratories in the MQI Research funding: None declared.
project; Employment or leadership: None declared.
2) involving Accreditation Bodies, so that the MQI may Honorarium: None declared.
be identified as a suitable tool complying with the ISO Competing interests: The funding organization(s) played
15189:2012 requirements; no role in the study design; in the collection, analysis, and
3) selecting and nominating a National Leader to coordi- interpretation of data; in the writing of the report; or in the
nate and manage the MQI project; decision to submit the report for publication.
4) defining guidelines supporting the use of QIs and
implementation of improvement actions in clinical
laboratories;
5) establishing criteria to ensure that an appropriate list
References
of QIs (number, typology, and frequency of collection 1. ISO Guide 73:2009. Risk management – Vocabulary. Geneva,
of data) is included in the MQI, procedures are pro- Switzerland: International Organization for Standardization,
cessed and laboratory performance evaluated; 2009.
6) sharing QIs with other inter-laboratory quality man- 2. Remona Eliza D, Monodora D. Risk management in laboratory
agement providers. medicine: from theory to practice. Acta Medica Marisiensis
2015;61:327–7.
3. Sciacovelli L, Sonntag O, Padoan A, Zambon CF, Carraro P, Plebani
M. Monitoring quality indicators in laboratory medicine does not

Conclusions automatically result in quality improvement. Clin Chem Lab Med

2011;50:463–9.
4. Shcolnik W, de Oliveira CA, de Sao Josè AS, de Oliveira Galoro CA,
The increased value of laboratory information as well Plebani M, Burnett D. Brazilian laboratory indicators program.
as its impact on clinical outcomes is a catalyst in assur- Clin Chem Lab Med 2012;50:1923–34.
ing the reliability of test results. The inter-laboratory 5. Kirchner MJ, Funes VA, Adzet CB, Clar MV, Escuer MI, Girona JM,
comparison of QIs is an important component of quality et al. Quality indicators and specifications for key processes in
clinical laboratories: a preliminary experience. Clin Chem Lab
management, since it enables a direct comparison with
Med 2007;45:672–7.
both other laboratories and established performance 6. Barth JH. Clinical quality indicators in laboratory medicine. Ann
specifications (benchmark). The QIs system, which Clin Biochem 2012;49:9–16.
 Sciacovelli et al.: Quality Indicators in Laboratory Medicine      357

7. Plebani M, Astion ML, Barth JH, Chen W, de Oliveira Galoro CA, 11. Plebani M, Sciacovelli L, Lippi G. Quality indicators for laboratory
Escuer MI, et al. Harmonization of quality indicators in labora- diagnostics: consensus is needed. Ann Clin Biochem 2011;48:479.
tory medicine. A preliminary consensus. Clin Chem Lab Med 12. Sciacovelli L, Aita A, Padoan A, Pelloso M, Antonelli G, Piva E,
2014;52:951–8. et al. Performance criteria and quality indicators for the
8. ISO 15189:2012. Medical laboratories – requirements for quality ­post-analytical phase. Clin Chem Lab Med. 2016;54:1169–76.
and competence. Geneva, Switzerland: International Organiza- 13. Plebani M, Sciacovelli L, Aita A, Pelloso M, Chiozza ML. Per-
tion for Standardization, 2012. formance criteria and quality indicators for the pre-analytical
9. Sciacovelli L, Plebani M. The IFCC Working Group on laboratory phase. Clin Chem Lab Med 2015;53:943–8. Erratum in: Clin
errors and patient safety. Clin Chim Acta 2009;404:79–85. Chem Lab Med 2015;53:1653.
10. Sciacovelli L, O’Kane M, Skaik YA, Caciagli P, Pellegrini C, 14. Westgard JO. Six sigma quality, design and control. Madison,
Da Rin G, et al. Quality indicators in laboratory medicine: from WI: Westgard QC, 2006.
theory to practice. Preliminary data from the IFCC Working 15. Lippi G, Bonelli P, Rossi R, Bardi M, Aloe R, Caleffi A, et al. Devel-
Group Project “Laboratory Errors and Patient Safety”. Clin Chem opment of a preanalytical errors recording software. Biochem
Lab Med 2011;49:835–44. Med 2010;20:90–5.