UK NEQAS FOR ANTIBIOTIC ASSAYS

THE ANNUAL REPORT 2008-2009

Department of Medical Microbiology

Southmead Hospital

Westbury-on-Trym

Bristol

BS10 5NB

Scheme Staff

The staff responsible for ensuring the running of the scheme for the distribution year 2008-09 were:

Scheme Organiser Professor Alasdair MacGowan

Scheme Manager Mrs Rhiannon John / Mr Alan Noel

Distribution Manager Mr Alan Noel / Dr Mervyn Darville

Scheme Packers Mrs Hazel Bennett / Mr Terry Fernandes / Dr Mervyn Darville

Mrs Rhiannon John commenced her maternity leave in Novemver 2008 and so Mr Alan Noel succeeded her as Scheme Manager with Dr Mervyn Darville performing some of the duties as acting Distribution Manager. Mr Terry Fernandez retired in January 2009. We thank him for all his efforts and wish him well in his retirement.

Scheme Contact Details

You can get in touch with the scheme by post, email and telephone:

UK NEQAS for Antibiotic Assays

Department of Medical Microbiology

Southmead Hospital

Westbury-on-Trym

Bristol, BS1 5NB

Tel: 0117 9596214 (answer phone service)

Fax: 0117 9593217

Email: ukneqas.antibiotics@nbt.nhs.uk

Scheme Introduction

This scheme provides a quality assessment service in antibiotic therapeutic drug monitoring to allow laboratories to monitor their performance, from drug assay through to submission of results. Twelve distributions are circulated annually. All participants can take any combination of seven antibiotics in ~900ul serum:

Amikacin, Gentamicin*, Flucytosine, Netilmicin, Teicoplanin, Tobramycin and Vancomycin*

*Gentamicin/Vancomycin are distributed as a combined sample.

There was no change in the antibiotics offered this year, and the number of participants taking netilmicin and flucytosine remained the same. As in previous years flucytosine and netilmicin were not scored due to the low number of participants.

Registered Participants

This year our scheme continued to experience some loss of participants, mostly down to the merger of pathology laboratories and the loss of antibiotic assay testing in many microbiology laboratories within the UK.

Analyte	No. of Participating Laboratories
Amikacin	81
Flucytosine	5
Gentamicin	208
Netilmicin	3
Teicoplanin	25
Tobramycin	104
Vancomycin	205

CPA Accreditation

The scheme was inspected January 2008, and retained accreditation until January 2010.

Scheme Developments

We are committed within the Antibiotic Assay scheme to develop and implement new, more-up-to-date software and IT in order to streamline the distribution process, both for staff and participants. The first stage of IT development took place in 2007 with the introduction of the new web-based Wolfson software in April 2007. The new look participant reports met with approval, and all of you seemed to agree that they were much easier to read and interpret than the old style of form.

IT development continued during this distribution year with the implementation of online results entry and online report download. Currently around 33% of laboratories returned results online.

The educational reporting scheme that would run alongside the main antibiotic assay scheme is still currently under review. This educational resource is becoming more widely available and it is important that the scheme remains current in the ideas surrounding other EQA schemes. Almost all participants who replied to our questionnaire were very interested in the pilot scheme.

Scheme Audits and Performance Targets

Table 1: Communication audit for the distribution year

Communication Type	Number (Percentage)
Routine queries	127 (56.6%)
Enrolment queries	25 (11.2%)
Repeat samples	23 (10.3%)
Scoring queries	22 (9.8%)
Complaints	14 (6.3%)
Invoice queries	13 (5.8%)
Total	224

Complaints

There were 14 complaints in total. 5 were samples lost in transit. 4 were samples sent to incorrect laboratories. 2 were regarding the quality of the samples (serum issues).

Fax problems

If possible, please remember that in the few working days prior to the distribution closing date the fax machine is extremely busy and it may take several minutes for your fax to go through correctly.

If you have problems with faxing then alternatively, email your results with “Results return–and the distribution number” in the email header. We advise all laboratories to use the web-reporting of results. The return of results by fax or email can be easily audited, and will avoid loss of returns in the postal system.

Packaging errors

We had 4 packaging errors during the distribution year – one incorrect sample was sent to a laboratory and one laboratory received the wrong paperwork.

Specimen quality

Issues with serum quality has meant that our scheme has begun trying to source alternative supplies of serum. We now purchase our serum from the same source as the HEATH Control EQA Scheme. This commenced from September 2008 and we have subsequently had no issues arising from the quality of the specimens.

Distribution Year Statistics

In order to understand the statistics provided with this report it may be helpful to read the following information first. A lengthier version of the scoring system is available online in the Participants Manual (www.ukneqasaa.win-uk.net)

How the Scheme is Scored

Method and sub-method performance

The performance of each method is summarised. Performance can be assessed by looking at the ALTM’s - All Laboratory Trimmed Mean and the percentage coefficients of variation (CV%’s), which are calculated in the following way

Coefficient of variation = standard deviation / mean x 100

Statistical trimming results in the removal of mistakes and inaccurate results from the calculations. The presented data allows laboratories to compare the reproducibility of different methods and of sub-methods. However, do be cautious when comparing results from a method with a large number of participants to one with only few.

Laboratory methodology performance

Each month a laboratory measures the concentration of the quality control sample. A % bias calculation is performed in order to assess the extent of the error it has made shown as the %error.

% Bias = (laboratory result – target concentration)/ target result x 100

Bias can be positive (consistently above the target value), negative (consistently below the target value), fixed (always x mg L^-1), proportional (always x%) or variable.

This result is added to the previous five results and the modulus mean % error + 2.S.D. (MEAN+2SD) is calculated. This value is related directly to the score and to the grouping and thus the laboratory’s performance rating.

Please refer to The Participant’s Manual (available on-line, through our web-site www.ukneqasaa.win-uk.net) to see how the groups and scoring scheme are linked. In summary:

A score of 2 indicates an excellent performance

A score of 1 indicates a borderline performance or may be indicative of a future problem

A score of 0 to -1 indicates a poor performance.

Please note that blunders (mistakes such as transcription errors or the mixing up of results) are not excluded when determining laboratory performance. These are the commonest errors and are made every month. Even including blunders, the proportion of poor performers is small and true persistent poor performance is rare.

Causes of poor performance and the potential clinical implications

For virtually all laboratories, assay performance is always very good. The number of poor performers remains small. Blunders are common but are not repeated every month by a single laboratory. One main cause of poor performance is made by submitting a single transcription error. A few mistakes are also made by the transposition of two antibiotic assay results which will then result in a poor score for two antibiotics for a number of months. Whilst noting that such transcription/transposition errors will not have any adverse clinical impact, they are not mistakes that can be ignored.

Recently we have had received some poor quality faxed returns – numbers written too faintly or with lines obscuring the results. This has meant that incorrect results were entered that then resulted in a poor score. We do not know what proportion of our members fax their clinical results but clearly there is a potential cause of serious error if a fax is not clear. Since the commencement of the web-reporting option for result return, the number of faxes received and therefore the number of poor quality faxed returns received is decreasing.

Another commonly made error is the failure to multiply-up results when samples have required dilution to bring them within assay range. Clearly, the failure to multiply-up the results for clinical samples would be a serious error. Very occasionally a laboratory may assay the wrong samples. These may be a previous month’s circulation, internal quality samples or actual patient samples. All of these types of mistakes are potentially serious and such laboratories should scrutinise and tighten their procedures to ensure this cannot happen with clinical samples.

Once again, it must be remembered that where the numbers of participants are low, comparisons should be made with caution. This is even more important when comparing mean scores than it is to method CV%s. One poor performer in a small group will have considerably more impact on the mean score than it will in a large group.

Method Codes

This year has seen a considerable increase in the number of methods being used by participants. The current list of method and sub-method codes is shown in the tables below with the number of participants. Only methods with more than 13 participants are scored as a peer group. In a change to our report forms participants can only see methods with more than 13 participants (unless your lab happens to use a small method in which case the method results will appear).

Table 2: Method and Sub-method Codes

Manufacturer	Method	Method code	Sub-method	Sub-method code
Abbott	TDx / FLx	TDX	Biostat PFIA	BIO PFIA
Abbott	Abbott	ABBOTT	Architect	ARCH
Abbott	Abbott	ABBOTT	Axsym	AXM
Beckman	Beckman	BECKMAN		DXC
Behring	Emit	EMIT	Wet manual Auto analyser	WM AA
Behring	Petina	PETINA
Bioassay	Gram negative	GMNEG
Bioassay	Gram positive	GMPOS
Bioassay	Yeast	YEAST
Biostat	Biostat	BIOSTAT	Biostat PFIA Seradyn	BIO PFIA SER
Boehringer Mannheim	Cedia	CEDIA
Dade	Dade	DADE
Launch	Biokit	BKT
Olympus	AU400/600/640 2700/5400	OLYMPUS
Plate assay	Plate assay	PLATE
Roche	Roche	ROCHE	PFIA	PFIA
Roche	Modular	ROCHEMP
Siemens	Siemens	SIEMENS	Centaur	CENT
Siemens	Siemens	SIEMENS	Advia	ADVIA
Siemens	Siemens	SIEMENS	Technicon	TECH
Various	HPLC
Unknown	No method

The graphs below show the breakdown of all methods used by our participants by antibiotic. The majority of participants still use the TDx or FLx for amikacin, gentamicin, tobramycin and vancomycin. However, there has been a steady change in methodology away from the TDx, particularly for vancomycin where the Abbott Architect is now nearly as big a method group.

Antibiotics and Methods for Distribution Year 2008 - 2009

Distribution Results

2 analytes were not scored over the distribution year, an amikacin and a gentamicin. The target value will not appear on your written report if the analyte has not been scored.

Table 3 Participant Returns Per Distribution

DIST	ANAL	Parts	No.	Non-Rets	W.I.	ALTM	SD	CV
2300	AMIK	111	95	12	56	53.9	4.73	8.76
2299	AMIK	111	100	10	18	17.7	1.04	5.87
2298	AMIK	113	98	12	22	19.9	1.18	5.92
2297	AMIK	114	94	19	4.8	4.4	0.35	7.98
2296	AMIK	115	101	12	20.5	18.4	0.99	5.39
2295	AMIK	115	92	21	30	26.3	1.63	6.19
2294	AMIK	115	96	16	13.8	12.7	0.77	6.09
2293	AMIK	114	96	12	57	50.7	4.13	8.15
2292	AMIK	114	97	15	15.6	14.4	0.95	6.56
2291	AMIK	115	98	16	25	22.2	1.2	5.39
2290	AMIK	115	97	13	2.5	2.7	0.38	14.2
2289	AMIK	115	100	13	20	20.1	1.01	5.03
2300	FLU	10	8	1	48	47.4	3.57	7.52
2299	FLU	10	8	1	132	121.7	18.91	15.55
2298	FLU	10	7	2	62	62.8	7.74	12.32
2297	FLU	10	9	1	77	65.7	9.7	14.76
2296	FLU	10	7	2	36	25.4	11.5	45.28
2295	FLU	10	8	2	63.5	61.8	4.36	7.06
2294	FLU	10	8	2	92	72	21.13	29.33
2293	FLU	10	9	1	51	48.6	6.68	13.76
2292	FLU	10	8	2	118	109.4	17.14	15.67
2291	FLU	10	7	3	71	67.6	7.58	11.22
2290	FLU	10	8	2	80	78	7.57	9.71
2289	FLU	10	9	1	33	30.3	3.53	11.66
2300	GENT	275	246	23	5	4.6	0.3	6.4
2299	GENT	276	246	20	10.3	9.4	0.69	7.31
2298	GENT	291	197	36	0.3	0.4	0.12	28.21
2297	GENT	289	243	39	9.2	8.5	0.58	6.76
2296	GENT	290	250	26	12.1	11	1.14	10.41
2295	GENT	290	248	36	3.1	2.9	0.18	6.07
2294	GENT	290	238	30	15.4	14.3	1.16	8.15
2293	GENT	291	261	25	6.7	6.2	0.6	9.74
2292	GENT	292	249	34	10.5	9.4	0.76	8.17
2291	GENT	292	255	31	1.4	1.3	0.17	13.16
2290	GENT	292	253	30	7.9	7.2	0.48	6.65
2289	GENT	292	252	27	12.3	11	1.17	10.59
2300	NET	13	5	4	2.1	2.3	0.15	6.71
2299	NET	13	4	4	9.2	7.9	0.93	11.74
2298	NET	15	5	5	12.5	14.6	7.33	50.23
2297	NET	15	4	6	3.1	2.5	0.51	19.87
2296	NET	15	5	4	15.8	12.7	4.14	32.67
2295	NET	13	4	3	7.2	6.9	0.74	10.72
2294	NET	14	5	3	11.3	10.2	0.97	9.5
2293	NET	14	4	5	2.4	2.6	0.41	15.97
2292	NET	14	6	3	7.5	7.8	1.7	21.88
2291	NET	14	6	3	13.7	12.6	4.58	36.45
2290	NET	14	6	4	4	4	0.81	20.06
2289	NET	14	8	4	16.1	15.1	1.43	9.47
2300	TEIC	30	27	3	21.2	17.8	1.58	8.87
2299	TEIC	29	26	3	8.4	8.7	1.03	11.77
2298	TEIC	29	25	4	30.1	29.2	1.77	6.07
2297	TEIC	27	25	2	58.4	55.8	4.94	8.85
2296	TEIC	27	24	3	18.4	18.3	1.48	8.08
2295	TEIC	27	23	4	8.9	9.3	0.94	10.21
2294	TEIC	27	24	2	45.6	45.5	3.55	7.8
2293	TEIC	27	22	3	3	2.9	0.73	25.54
2292	TEIC	27	25	2	18.9	19.2	1.46	7.61
2291	TEIC	27	23	3	61.2	59.7	3.94	6.6
2290	TEIC	27	25	2	13.6	14.3	1.03	7.23
2289	TEIC	28	25	2	41.2	43	4.16	9.68
2300	TOB	141	121	13	11.2	10.6	0.99	9.38
2299	TOB	141	125	12	1.2	1.2	0.19	15.36
2298	TOB	143	122	16	7.6	7.7	0.66	8.6
2297	TOB	141	114	16	12.4	12.4	1.2	9.71
2296	TOB	141	121	13	4.4	4.4	0.45	10.16
2295	TOB	139	114	17	14.9	16.5	1.54	9.34
2294	TOB	138	116	16	6.4	6.5	0.53	8.07
2293	TOB	137	116	14	10.1	10.5	1.04	9.93
2292	TOB	136	116	14	1.8	1.9	0.18	9.33
2291	TOB	138	118	14	9	9.4	0.87	9.3
2290	TOB	138	117	14	12.2	12.4	1.4	11.3
2289	TOB	138	122	10	3.5	3.5	0.29	8.37
2300	VANC	276	243	24	50	49.4	4.5	9.11
2299	VANC	276	250	18	13.5	14	1.02	7.3
2298	VANC	304	246	46	67	66.5	5.94	8.92
2297	VANC	303	247	50	21.3	22.3	1.76	7.9
2296	VANC	303	255	39	40	40.2	3.75	9.34
2295	VANC	302	248	49	6.6	6.7	0.75	11.29
2294	VANC	300	250	44	28.6	28.8	2.31	8.01
2293	VANC	301	257	39	44	43.6	3.49	7.99
2292	VANC	302	253	43	15.1	15.4	1.16	7.53
2291	VANC	302	254	41	60	59.4	5.77	9.72
2290	VANC	302	253	42	23.5	22.8	1.88	8.25
2289	VANC	302	255	38	38.9	24.7	2.02	8.18

Repeat Samples

The table below shows the number of repeat samples sent to participants with an error of more than 30% over the distribution year.

	AMIK	FLUC	GENT	NETIL	TEICO	TOBRA	VANC
Apr			2		1		2
May					1	1	2
Jun			8				2
Jul			3			1	1
Aug	1		4			3	5
Sep	1		4		1		1
Oct						2	7
Nov			3		1	1	2
Dec			1			2	2
Jan						1	1
Feb	1		1			3	1
Mar	1		1		1	1	1
TOTAL	4	0	27	0	5	15	27

Poor Performing Laboratories

Biannually, the poorly performing laboratories are analysed (in March and September). Laboratory numbers have been removed and replaced with letters. The same coding system is used for both months, but each summary contains the results for different laboratories. Only laboratories in the United Kingdom AND Northern Ireland are highlighted as poor performers.

LID	ANAL	METHOD	SMETHOD	COUNTRY	LAB_TYPE	Mean+2SD	Score
A	AMIK	TDX	PFIA	UK	NHS	53.9	0
B	GENT	ABBOTT.	ARCH	UK	NHS	55.9	0
C	AMIK	DADE		UK	NHS	97.3	0
D	GENT	BIOSTAT	SER	UK	NHS	81.9	0
E	VANC	ABBOTT.	ARCH	UK	NHS	70.7	0
F	GENT	SIEMENS	DIMENSION	UK	NHS	540.9	-1
F	VANC	SIEMENS	DIMENSION	UK	NHS	84.1	0
G	GENT	BIOSTAT	SER	UK	NHS	125.3	-1
H	VANC	TDX	PFIA	NI	NHS	57.8	0
I	TOB	BIOSTAT	SER	UK	NHS	50.6	0
J	GENT	ABBOTT.	ARCH	NI		153.4	-1

September 2008

LID	ANAL	METHOD	SMETHOD	COUNTRY	LAB_TYPE	Mean+2SD	Score
A	TOB	ABBOTT.	ARCH	UK	PHL	70.2	0
B	AMIK	TDX	PFIA	UK	PHL	64.8	0
C	VANC	TDX	PFIA	UK	NHS	52.4	0
D	VANC	TDX	PFIA	UK	NHS	52	0
E	TOB	BECKMAN	DXC	UK	NHS	120.2	-1
F	GENT	ROCHE	PFIA	UK	NHS	76.7	0
F	VANC	ROCHE	PFIA	UK	NHS	101.6	-1
G	GENT	ROCHEMP		UK	NHS	113.2	-1
G	TOB	ROCHEMP		UK	NHS	66.7	0

March 2009

End of Annual Report 2008 -2009