Read this article in:

"Ghost sows" and the other 8 most frequent errors when processing data (2/2)

This second part deals with the problems caused by not  recording properly the cause of abortions, sows and piglets deaths, nurse sows and hormonal treatments.

5. Abortions without a specific cause. Is one of the most frequent and more serious errors, since its impact on the analysis of abortions is very important. As a general rule, the first proven event must be always recorded (i.e., a sow that returns to heat because it's just had an abortion must be recorded as "abortion"; a sow that has an abortion due to an injury must be recorded as "injury / accident"; a sow with a negative diagnosis of gestation due to metritis must be recorded as "metritis", etc). This is particularly common after 40 days of gestation, when a period of remarkable darkness begins.

6. Sows culled without a cause or with a too generic cause. Causes are too often recorded as "other causes" or "underachiever". A little effort invested in both standardizing the causes and then precisely recording the reason for each elimination, will allow us to obtain high quality information, which is quite unusual.

  Cause of death
  No. Weaned not serviced Abortion Barren Pregnant Gilt not serviced
  No. % No. % No. % No. % No. %
R00 reproductive problem 19 3 15.88 0   0   15 78.9 1 5.3
R03 Abortion 1 0   0   0   1 100.0 0  
X00 Other causes 38 9 23.77 0   0   28 73.7 1 2.6
Total 58 12 20.7 0   0   44 75.9 2 3.4

Figure 3. Farm with a 66% of sales of sows without a recorded cause.


7. Pre-weaning mortality without dates and with confusing causes. 

In most cases, the operator's impression when the veterinarian asks "What is the main cause of mortality in piglets and their age?" does not correspond to reality.

This is because some causes stand out more than others to the operators' eyes and, above all, because they quickly get used to what they see every day on the farm. When this happens, abnormal events fail to draw the workers' attention. It is therefore essential to build upon objective data in order to obtain objective information and make appropriate decisions.

Moreover, misconceptions were found in many cases, such as the inability to clearly differentiate between an "under-developed" animal and a "non-viable" one. The root causes are different: lack of milk and feeding problems during gestation, respectively. Therefore, there should never be older than 4-5 days "non-viables".

In order to determine the normal levels within each age group, the type of genetics present in the holding must also be taken into account. As Figure 4 shows, the distribution pattern has changed over the years with the increased presence of this type of genetics, decreasing peripartum mortality and moving to the next age group (70% in 2006, whereas in 2015 it was 41%.)

Evolution of the distribution of pre-weaning mortality  depending on the age at death in the last ten years.

Figure 4. Evolution of the distribution of pre-weaning mortality  (PWM) depending on the age at death in the last ten years.
Based on data from 3,102,130 services


Pre-weaning mortality by cause of death. The graph shows a farm that records deaths incorrectly.

Figure 5. Pre-weaning mortality by cause of death. The graph shows a farm that records deaths incorrectly.

Pre-weaning mortality by cause of death

Figure 6. Pre-weaning mortality by cause of death. The graph shows a farm that records deaths correctly.


8. Inadequate processing of nurse sows.

Hyper-prolific genetics increasingly require more and higher-quality work in the farrowing unit, so failure to record this and ignoring its consequences (for better or worse) are both unacceptable for a professional producer. Obviously, a software that can properly process this kind of management (which is not easy and therefore not common) must be available in the first place.

When these data are available, they can be analysed for indicators such as the difference between weaned per litter and weaned per sow, parameter showing the efficiency of the use of nurse sows. The number of weaned / litter indicates the average size of a weaned litter, whereas the number of weaned / sow takes into account the real extra weaned piglets through the use of partial weanings or nurse sows.

Sows completely weaned 105 118 119 98 110
Sows weaned 17 14 28 15 18
Sows weaned without a litter 0 4 4 2 2
Litters weaned 126 142 139 106 128
Use of a nurse sow 21 28 24 10 21
Weaned piglets 1617 1733 1727 1248 1581
(from partial weanings) 52 0 0 0 13
Average weaned/litter 12.8 12.2 12.4 11.8 12.3
(Less piglets from partial weanings) 12.4 12.2 12.4 11.8 12.2
Average weaned/sow 15.4 14.7 14.5 12.7 14.4
(Less piglets from partial weanings) 14.9 14.7 14.5 12.7 14.3

Figure 7. Weaning parameters of a farm where nurse sows and partial weanings are recorded.


9. Ignoring treatments. Hormone products (gonadotropins, altrenogest, prostaglandins, vetrabutine or oxytocin) are regularly used to adjust heats, synchronize, facilitate or speed up farrowings, but rarely recorded and processed in order to determine their impact on the variable that needs to be controlled (onset of heat, farrowing day, milk production, etc.) It is a remarkable mistake, since it limits decision-making ability in a critical area on the farm, both for its technical and economic impact.

Date Farrowing Event Details
26-Oct-15   Dead piglet 1 crushed (4 days old)
31-Oct-15   Transfer 1 piglet to sow 242
19-Nov-15   Total weaning 10 piglets (2 piglets less) (28 days)
18-Dec-15   X1 1st service Semen: PIETRAIN (PM) (29 days)
18-Dec-15   X2 Service Semen: PIETRAIN (AM)
19-Dec-15   X3 Service Semen: PIETRAIN (AM)
20-Dec-15   X4 Service Semen: PIETRAIN (AM)
11-Apr-16 8 Farrowing 6 born alive, 2 stillbirths, 0 mummified (115 days)
11-Apr-16   Adoption 2 piglets from 136
11-Apr-16   Adoption 2 piglets from 348
15-Apr-16   Transfer 1 piglet to 343
05-May-16   Total weaning 9 piglets (24 days)
16-May-16   X1 1st service Semen: PIETRAIN (PM) (11 days)
16-May-16   X2 Service Semen: PIETRAIN (AM)
16-May-16   Mark G-PG 600
17-May-16   X3 Service Semen: PIETRAIN (AM)

Figure 8. Recorded history of a sow treated with PG 600.

All errors described are very common in most farms. The accumulation of such errors hinders the work of technical and farm managers, as reports or worklists are implemented with incorrect or unreal data.

We recommend to correct two errors per month which, within a gestation period, will allow you to considerably increase your ability to improve your production.

Article Comments

This area is not intended to be a place to consult authors about their articles, but rather a place for open discussion among pig333.com users.

Allan Schinckel 15-Nov-2016 (11 months 3 days ago)

good article - mistakes that are easily done - and impact the records.

Access restricted to 333 users. In order to post a comment you must be logged in.

Not a registered user of 333?sign upand access swine prices, the search engine, ...
It is fast and free
Are you registered in 333?LOGINIf you've forgotten your password we'll send it to you here