Animals and housing
The study was conducted in five commercial Swedish pig farms from November 2015 to June 2017. All participation in the study was voluntary, and no compensation was offered to the farmers. All participating farms had commercial pig production in mechanically ventilated and insulated buildings and received veterinary advice regularly according to Swedish regulations. Specific criteria for participation in the study was daily straw provision and partly slatted flooring systems. Due to practical reasons, all farms were located in the south west of Sweden and farms were identified in collaboration with the local farm veterinary advisory service. All pigs were progenies of crossbred sows (Landrace and Yorkshire sows; either Norwegian Landrace*Swedish Yorkshire or Norwegian Landrace*Topigs Large White, that were inseminated with Duroc (G1, G2, F2, F3, G5, F5) or Hampshire boars (F4).
In order to investigate effect of age/size of the pigs, the experiment was conducted on both growers and finishers. One batch each of growing pigs (10 to 30 kg) was investigated on three grower farms: one specialized piglet producing farm (G1), and two farrow-to-finish farms (G2, G4). One batch each of finishing pigs (30 kg – 120 kg) was investigated on four farrow-to finish farms (F2, F3, F4, F5). On F2 and F4 the same batch of pigs was observed in the grower finishing pig stable. The experiment started as the pigs moved in to the grower/finisher unit and finished as the pigs were moved from the grower unit to the first pig was sent to slaughter, respectively. For practical reasons, it was impossible to conduct the experiments at the same period in all farms, and hence the season varied between farms and batches (Table 1). Detailed information about farms and housing is found in Table 2.
The pigs had daily straw provision (wheat straw, cut in all farms, except Farm 2, see Table 2) and supervision according to normal farm routines. The pigs in each batch were mixed in pens with gilts and castrates and sorted by size; housing the heaviest and the lightest pigs together respectively. Intact litters was practiced when possible, i.e. when the number of pigs in the litter matched the number of places in the pen and the size of the pigs were uniform within the litter. Whenever intact litters were not possible, pigs were sorted by size as described above. All pigs were undocked (according to Swedish legislation). This study was part of a larger study investigating the impact of increased straw ration on pig behaviour and prevalence of tail lesions (see ).
Experimental design and treatments
One batch of pigs, raised within the same physical unit, were studied on each farm. Pens deviating from the average pen regarding pen design or number of pigs were excluded from the study. In Farm 1, only half of the batch was, due to practical reasons, included in the study. In Farm 4 one row of pens in the grower stable was excluded from the study due to deviating climatic circumstances. In Farm 2 and 4 pigs were first studied in the grower stable and then followed into the finishing pig stable.
The studied pens were divided in two equally sized treatments per farm: Control (C), receiving the farm normal straw ration; and Extra Straw (ES), receiving a double C-ration (Table 2). The C-ration was determined and standardized before the experiment by measuring the daily straw ration provided by the herdsmen (Table 2). All pens of the same treatment were located in the same row of the stable unit, to ease for the animal caretaker. Except for straw ration, all pens in the stable unit were on farm level, managed in the same way. Any exceptions from the straw provision routine were, recorded by the caretaker (Table 3). If there was blockage in the slatted flooring, and more than 50% of the slatted area was no longer visible, the daily straw provision could be paused until blockage was cleared.
The pig and pen hygiene was scored every second week, including the first and last week of the experiment by two observers (not the farmer) (taking turns scoring C or ES treatment, no intra-observer reliability was measured) (Table 3). The recordings were conducted at least one hour after the daily cleaning and straw provision.
Pig hygiene was scored according to the Welfare Quality® protocol applied to growing and finishing pigs manure on the body (Welfare Quality 2009). All pigs were individually assessed for manure on the body, on the side that was visible towards the observer according to a three point scale; 1 if a maximum of 20% of the pig was covered in manure; 2: >20-50% of manure coverage; 3: >50 of manure coverage.
The solid and slatted floor of the pen were each divided in to four separate parts (Figure 1). The solid floor part was considered dirty when at least 50% of the area was covered by faeces, mired straw or were wet. The slatted floor area was considered as blocked, when the slats was covered and no space between the slats was visible for at least 50% of the assessed area. For each part assessed as dirty/blocked, the pen was scored with one point. Subsequently the scores were added to receive the final hygiene score of the solid and slatted area respectively. A pen could have a maximum of 4 points (all parts ≥50% dirty/blocked) and a minimum of 0 points (all parts <50% dirty/blocked) per solid/slatted area.
All data were recorded in Microsoft Excel 2016 and analysed through SAS 9.4 (SAS Institute Inc., Cary, NC, USA). The descriptive statistics were calculated through means and frequencies on both farm and age category (grower/finisher) level. Pig and pen hygiene data was ordinal and data had repeated measurements on pen level.
The low variability made analysis of variance on age level impossible. To enable analysis the data was rearranged into binomial traits; hygiene score 1 (i.e. maximum 20% soiled body surface) were considered clean while pigs with score 2-3 (>20% soiled body surface) were considered dirty. Data was then analysed using Fisher’s exact test, investigating impact of Treatment (C or ES) on Pig hygiene on pig level for each Farm and observation occasion separately.
The low variability made analysis of variance on age level impossible. To enable analysis of variation, the data was rearranged into binomial traits on the solid/slatted floor separately. Pens with the score 0 (no soling/blockage of the floor) were considered clean while pens with score 1-4 (soling/blockage of 25-100% of the Solid/Slatted area) were considered dirty. Data was then analysed using Fisher’s exact test, investigating the impact of Treatment (C or ES) on pen hygiene on pen level for each Farm and observation occasion separately.
Correlation between Pig and Pen hygiene
Spearman rank correlation was used to investigate the correlation between pig hygiene and solid/slatted floor hygiene to investigate the hypothesis if pig and pen hygiene was correlated. Pig hygiene was converted into a mean score per pen instead of the initial pig level value. Solid and slatted hygiene was kept as the initial values ranging from 0-4 for each occasion. Thus, the correlations were estimated based on pen level scores, for each scoring occasion, both on Farm and Treatment level.