ANNEX I

Technical specifications referred to in Articles 3(1), 4 and 5(2)(b)

1.   OVERVIEW OF THE SURVEY

The survey shall be performed according to the overview in Figure 1.

Figure 1:   overview of the survey

x ‘breeding holdings’, 10 routine samples per holding (Part 1)

Survey

Selection of 10 holdings for estimation of the within-holding prevalence by taking 100 individual samples

y ‘production holdings’, 10 routine samples per holding (Part 2)

2.   SAMPLING FRAME

2.1.   The delineation of the population

The survey shall be carried out on holdings harbouring at least 80 % of the breeding pig population in a Member State. Preferentially holdings having 50 breeding pigs or more shall be sampled. However, if those holdings having 50 breeding pigs or more do not contain 80 % of the national herd of breeding pigs, then smaller holdings with less than 50 breeding pigs shall also be sampled.

Holdings with breeding pigs shall be classified either as ‘breeding holdings’ or as ‘production holdings’. Breeding holdings sell gilts and/or boars for breeding purposes. Typically, they sell 40 % or more of the gilts that they rear for breeding whilst the remainder are sold for slaughter. In contrast, production holdings mainly sell pigs for fattening or slaughter.

The Salmonella prevalence must be measured separately in breeding holdings (part 1 of the survey) and in production holdings (part 2 of the survey), representing the herds as indicated in Figure 2, but excluding weaner to finisher herds.

Figure 2:   overview of holdings

Nucleus or pure breeding herds

Breeder or multiplier or supplier herds (cross-breeds)

Farrow to finish herds

Farrow to weaner (or to grower) herds

Weaner to finisher herds

Breeding

Production

2.2.   The sample and the sampling strategy

Both parts of the survey shall have a similar two-stage sampling design. In the first stage, a random sample of holdings shall be selected in every Member State from the breeding holdings and a second random sample shall be selected from the production holdings group. The number of holdings required is discussed in section 2.3. In the second stage, a number of pens shall be selected for sampling within every selected holding (see section 2.2.2).

2.2.1.   First stage: selection of holdings

Each Member State must create two sampling frames. The first shall list all eligible breeding holdings (usually, those holdings with at least 50 breeding pigs — see section 2.1) and the second shall list all eligible production holdings. The required number of holdings for each part of the survey will then be selected at random from each list. A random sample is intended to ensure that the survey includes holdings with a range of herd sizes and from all regions of a Member State where pig keeping is practised. It is recognised that in some Member States, there may be a few holdings (e.g. fewer than 10 % of all eligible holdings) with a very large herd size. Thus, random selection may by chance result in none of these very large herds being selected. A Member State may use a stratification criterion prior to holding selection — for example, to define a stratum containing the 10 % of largest herds and to allocate 10 % of the required sample size to this stratum. Similarly, a Member State may stratify the sample across administrative regions according to the proportion of eligible herds within each region. Any stratification that is considered should be described in the report that a Member State submits to the Commission in accordance with paragraph 5(1).

If a selected holding cannot be sampled (for example, if it no longer exists when sampling is carried out) a new holding shall be selected at random from the same sampling frame. If any stratification (e.g. on herd size or region) was in operation, then the new holding should be selected from the same stratum.

The primary sample size (number of holdings to be sampled) shall be approximately equally distributed over the year to cover the different seasons, as far as possible. Samples shall be taken from approximately one twelfth of the number of holdings each month.

Outdoor holdings must be included in the survey but there shall be no mandatory stratification on this production type.

2.2.2.   Second stage: sampling on the holding

In each selected breeding herd and production herd the pens, yards or groups of breeding pigs over six months of age to be sampled will be randomly selected.

The number of pens, yards or groups to be sampled must be proportionally allocated according to the numbers of breeding pigs in the different stages of production (pregnant, non-pregnant and other categories of breeding pigs). The exact age categories to be sampled are not prescribed, but this information shall be collected during the sampling.

Breeding pigs that have arrived recently to the herd and are held in quarantine shall not be included in the survey.

2.3.   The sample size calculation

2.3.1.   Primary sample size (first-stage sample size)

A regular primary sample size calculation shall be conducted for the breeding holdings and a second regular primary sample size calculation shall be conducted for the production holdings. The primary sample size shall be the number of breeding holdings to be sampled and the number of production holdings to be sampled in each Member State and it shall be determined taking into account the following criteria, assuming simple random sampling:

(a)	the total number of breeding holdings (breeding holdings, part 1 of the survey);

(b)	the total number of production holdings (production holdings, part 2 of the survey);

(c)	annual expected prevalence (p): 50 %;

(d)	desired confidence level (Z): 95 %, corresponding to a Zα value of 1,96;

(e)	accuracy (L): 7,5 %;

(f)	using these values and the formula:

The calculation shall be conducted firstly, for the breeding holdings and secondly, for the production holdings. In each case, the assumptions in steps (c)-(e) above are the same.

For practical purposes, if there are 100 000 or more holdings in either the breeding herds sampling frame or the production herds sampling frame then that population can be considered to be infinite and the number of holdings to be randomly selected from that sampling frame is 171 (see Table 1). Where the number of breeding herds or production herds is less than 100 000 a finite population correction factor is applied and fewer holdings need to be sampled as shown in Table 1.

As an example, if there are in a Member State 1 000 holdings belonging to the production holding group and 250 holdings belonging to the breeding holding group, 147 holdings must be sampled in the production holding group and 102 holdings must be sampled in the breeding holding group.

Table 1

Number of holdings with breeding pigs to sample in either part of the survey as a function of the finite population size (total number of holdings with breeding pigs in the Member States)

Number of holdings with breeding pigs (N)	Sample size (n) for finite population size 7,5 % accuracy
100 000	171
10 000	169
5 000	166
2 000	158
1 000	147
500	128
250	102
150	80
125	73
100	64
90	59
80	55
70	50
60	45
50	39
40	33
30	26
20	18
10	10

Non-response shall be anticipated, e.g. by increasing the sample size in each group by 10 %. Any unsuitable holding shall be replaced in the process of the survey (see Section 2.2.1).

In case an estimation of the number of ‘breeding holdings’ is not possible prior to the start of the survey, a number of holding shall be selected for sampling as in Table 1 based on the total number of holdings with sows (X holdings). The number of holdings to be sampled shall be increased by at least 30 % ((X + 30 %) holdings). Prior to the survey, the competent authority shall identify a number of breeding holdings, equal to at least this additional 30 %. While visiting the farms, the holding will be classified as breeding or production holding according to the definitions above.

2.3.2.   Secondary sample size (second-stage sample size)

In each selected holding routine faecal samples (see section 3.1) shall be collected from 10 randomly chosen pens, yards or groups of breeding pigs. If necessary (for example, in farrowing accommodation or where sows are kept in small groups of less than 10 individuals), a group can consist of more than one pen. At least 10 individual breeding pigs should contribute to each routine sample.

However, where on small holdings or holdings with large numbers of breeding pigs kept outdoors in paddocks, the number of pens, yards or groups is less than 10, sampling of the same pen, yard or group shall be required so that a total of 10 routine samples are submitted.

3.   SAMPLE COLLECTION IN THE HERDS

3.1.   Type and detail of the routine sample

The material collected for bacteriological analysis shall be freshly voided faeces representing the whole holding, which is the unit of interest. Since every holding is unique, it shall be decided, before starting the sampling, which pens, yards or groups within the holding are sampled. The sample collected shall be placed in a separate container avoiding cross contamination and sent to the laboratory.

Each pooled sample shall total at least 25 g and two approaches may be employed to collect these pooled faeces samples:

(1)

where there is an accumulation of mixed faeces within an area of a pen or yard, a large swab (e.g. 20 cm × 20 cm) can be used to pass through the faecal mass, ensuring that at least 25 g of mixed material is collected. This may be achieved by for example, moving the swab along a two-metre zig-zag path such that it is well coated with faecal material. If necessary, for example due to hot weather or on slatted flooring then the swab may be moistened with an appropriate liquid such as drinking water;

(2)

where there is no such accumulation, for example in a field, a large yard, a farrowing house, or pens or other accommodation with low numbers of pigs per group, then individual pinches shall be selected from individual fresh faecal masses or places so that a minimum of 10 individuals, contribute to a total sample volume of at least 25 g. The sites from which these pinches are collected should be distributed in a representative manner across the area concerned.

Approach 1 shall be preferred where practical. In this approach at least 10 individual pigs must contribute to each sample taken, otherwise approach 2 shall be applied.

3.2.   Additional sampling for the within-holding prevalence study

Together, 10 holdings, selected at random from the overall sample of breeding holdings and of production holdings are subjected to more intensive sampling. On these holdings 10 routine samples shall be collected in the same manner as described previously (section 3.1). In addition, 10 individual samples of at least 30 g shall be collected in each selected pen and shall be identified in such a manner that these 10 individual samples can be associated with the routine sample from that pen. Thus in total, 10 routine samples and 100 (10 × 10) individual samples shall be collected from each of these 10 holdings. The processing of these samples is described in section 4.3.1.

This sampling should be applied in the Czech Republic, Denmark, Romania, Slovenia, Sweden and the United Kingdom.

3.3.   Sample information

All relevant information available from the sample shall be recorded on a sampling form produced by the competent authority to enable the data requirements in Part 5 to be fulfilled.

Each sample and its sample form shall be labelled with a unique number which shall be used from sampling to testing, and with the code of the pen. The competent authority must arrange for the issue and use of a unique numbering system.

3.4.   Transport of samples

Samples shall be preferably kept at between + 2 to + 8 °C and free of external contamination during transportation. The samples shall be sent to the laboratory as quick as possible within 36 hours by fast mail or courier and shall reach the laboratory no later than 72 hours after sampling.

4.   LABORATORY ANALYTICAL METHODS

4.1.   Laboratories

Analysis and serotyping shall take place at the National Reference Laboratory (NRL). Where the NRL does not have the capacity to perform all analyses or if it is not the laboratory that performs detection routinely, the competent authorities may decide to designate a limited number of other laboratories involved in official control of Salmonella to perform the analyses. These laboratories shall have proven experience of using the required detection method and have a quality assurance system complying with ISO standard 17025 and be submitted to the supervision of the NRL.

4.2.   Receipt of samples

At the laboratory, samples shall be kept refrigerated until bacteriological examination, which shall preferably be carried out within 24 hours after receipt but in any case no later than 96 hours after the sample was collected.

4.3.   Sample analysis

Member States shall guarantee that all involved parties have been sufficiently trained to carry out the analyses.

4.3.1.   Preparation

In the laboratory, routine samples shall be mixed carefully and thoroughly before 25 g are collected for analysis.

For evaluation of the within-holding prevalence in accordance with paragraph 3.2, each of the individual collected samples (30 g) needs to be divided into two parts. One part, weighing at least 25 g, shall be mixed carefully and thoroughly and subsequently cultured individually. The remaining second part is to be used to prepare an artificially pooled sample from the 10 individual samples in the selected pen, group or yard. This latter part shall be prepared by adding 10 times 2,5 g of the individual samples to create an artificially pooled sample of 25 g. The artificially pooled samples are mixed carefully and thoroughly before analysis. In total, 10 routine samples, 10 artificially pooled sample and 100 individual samples shall be analysed from each of the 10 holdings selected for the estimation of the within-holding prevalence.

4.3.2.   Detection and identification methods

4.3.2.1.   Detection of Salmonella

The method recommended by the Community Reference Laboratory (CRL) for Salmonella in Bilthoven, the Netherlands, shall be used. This method is described in the Annex D of ISO 6579: ‘Detection of Salmonella spp. in animal faeces and in samples of the primary production stage’. The latest version of Annex D shall be used.

4.3.2.2.   Serotyping of Salmonella

All strains isolated and confirmed as Salmonella spp. shall be serotyped according to the Kaufmann-White scheme, by the NRL for Salmonella.

For quality assurance, 16 typable strains and 16 non-typable isolates shall be sent to the CRL for Salmonella. A proportion of these isolates shall be sent to the CRL on a quarterly basis. If fewer strains have been isolated all shall be sent.

4.3.2.3.   Phage typing of Salmonella

In case isolates of Salmonella Enteritidis and Salmonella Typhimurium are phage typed (optional), the methods described by the WHO reference centre for phage typing of Salmonella of the Health Protection Agency (HPA) Colindale, London, shall be used.

5.   REPORTING FROM THE MEMBER STATES

5.1.   Overall description on the implementation of the survey

A report in text format shall include at least:

(a)	Member State;

(b)

description of the population of holdings with breeder pigs; (1) breeding holdings: (i) total number of breeder holdings; (ii) total number of nucleus holdings; (iii) total number of multiplier holdings; (iv) number of breeder holdings planned to be sampled, and number of breeder holdings actually sampled; number of holdings planned for sampling but not sampled and the reason therefor; (v) comments on the overall representativeness of the breeding holdings sampling programme; (2) production holdings: (i) total number of production holdings; (ii) total number of farrow to weaner/grower holdings; (iii) total number of farrow to finish holdings; (iv) number of production holdings planned to be sampled, and number of production holdings actually sampled; number of holdings planned for sampling but not sampled and the reason therefor; (v) possible comments on overall representativeness of the production holdings sampling programme;

(c)	number of samples obtained and analysed: (i) from breeding holdings; (ii) from production holdings; (iii) from holdings sampled for within-holding prevalence study;

(d)	overall results: (i) prevalence of breeding holdings and of production holdings infected with Salmonella, and serovars of Salmonella; (ii) outcome of within-holding prevalence study;

(e)	list of laboratories responsible in the baseline study for Salmonella: (i) detection; (ii) serotyping; (iii) phage typing (if carried out).

5.2.   Complete data on each holding sampled and corresponding tests results

The Member States shall submit the results of the investigation electronically in the form of raw data using a data dictionary and data collection forms provided by the Commission. This dictionary and forms shall be established by the Commission.

5.2.1.   The following information shall be collected in Member States for each holding selected for sampling:

(a)	code of the holding;

(b)	holding production type; (i) indoor versus ‘any stage of the production kept outdoors’; (ii) nucleus, multiplier, farrow to weaner, farrow to finish and farrow to grower;

(c)	holding size: the number of breeding pigs present at the time of sampling (adult inventory);

(d)	replacement policy: all replacement breeding pigs purchased; some replacement breeding pigs homebred or all replacement breeding pigs homebred;

(e)	(voluntarily:) clinical symptoms of diarrhoea: were there symptoms of diarrhoea within the three months before the sampling?

5.2.2.   The following information shall be collected in Member States for each sample sent to the laboratory:

(a)	code of the sample;

(b)	code of the laboratory involved in initial analysis;

(c)	date of sample collection;

(d)	date laboratory analysis begun;

(e)	detection of Salmonella: qualitative result (positive/negative);

(f)	serotyping of Salmonella: serovar(s) detected (may be more than one);

(g)	age of the pigs: all gilts versus mixed age breeding pigs;

(h)	sex: only sows; sows and boars or only boars;

(i)	production stage: maternity; mating, gestation (other?);

(j)	housing: slatted floor (fully/partly); solid floor; deep straw or other;

(k)	diet: are pigs in this pen, yard or group fed compound feed exclusively?;

(l)	feed supplement: is there any Salmonella reducing substance added to the feed (like organic acid, a probiotic)?;

(m)	systematic use of antibiotics: are antibiotics used in all animals of this group by any route of administration;

(n)	the last date of administration of antimicrobials to the animals (within the last four weeks).

5.2.3.   The following additional information shall be collected in Member States for each individual sample sent to the laboratory within the frame of the sampling for the within-holding prevalence:

(a)	code of the pooled sample;

(b)	detection of Salmonella in each individual sample: qualitative result (positive/negative);

(c)	serotyping of Salmonella in each individual sample: serovar(s) detected (may be more than one).

ANNEX II

Maximum Community financial contribution to the Member States referred to in Article 5

Member State	Maximum total amount for co-financing of analyses (EUR)
Belgium — BE	59 800
Bulgaria — BG	52 260
Czech republic — CZ	102 960
Denmark — DK	98 280
Germany — DE	57 980
Estonia — EE	9 360
Ireland — IE	43 420
Greece — EL	39 260
Spain — ES	82 680
France — FR	82 680
Italy — IT	79 300
Cyprus — CY	20 020
Latvia — LV	3 380
Lithuania — LT	13 780
Luxembourg — LU	11 960
Hungary — HU	74 360
Malta — MT	0
Netherlands — NL	87 100
Austria — AT	59 020
Poland — PL	85 020
Portugal — PT	54 860
Romania — RO	107 900
Slovenia — SI	81 120
Slovakia — SK	54 080
Finland — FI	64 740
Sweden — SE	81 120
United Kingdom — UK	102 960
Total	1 609 400

ANNEX III

Text of image