Cow characteristics (part one)

Each cow receives a parity (1 to 5+), which is based on a weighted herd average (CRV, 2014). Based on the parity, the cow receives a BW, a milk production level and a CI (CVB, 2012; CRV, 2014) (Table 1). Milk production level per parity is determined with a normal distribution function. The CI consists of a lactation period (350 to 361 days) and a dry period (60 days). A lactation curve is utilised to estimate the average daily milk production per cow based on weekly milk records (Wood, Reference Wood1967). Net energy requirement for maintenance, milk production, pregnancy and growth in case of parity 1 and 2 is estimated per day (CVB, 2012).

Table 1

Cow characteristics and probabilities of diseases per parity to estimate the economic impact of subclinical ketosis (SCK) in dairy cows

el=End of lactation period; Ad.=additional probability.

¹ Probabilities were obtained and calculated from data of the study Berge and Vertenten (Reference Berge and Vertenten2014).

² Additional probability of cow getting a disease with SCK were obtained and calculated from data of the study Berge and Vertenten (Reference Berge and Vertenten2014), probabilities were estimated as the difference between the probability of the cow getting the disease with SCK and the probability of the cow getting the disease without SCK.

³ Additional probability on getting culled or died were calculated for the disease.

⁴ Reduced milk production losses for displaced abomasum were estimated over 305 days but only occurred in months 1 to 4. Daily milk loss (kg/day) was estimated and reduced milk production losses occurred after day 30.

Energy requirement for maintenance was assumed to be 6.7% higher in summer (170 day/year) than in winter (195 day/year), because of grazing (CVB, 2012). The CI of all parities is longer than a year and therefore the summer and winter period of the CI was estimated using the yearly fraction of grazing (170/365). Finally, energy requirements and milk production of a cow are expressed in 365 days using the yearly fraction of the CI (365/CI).

Dynamics of subclinical ketosis and related diseases (part two)

The model includes SCK and related diseases mastitis, metritis, DA, lameness and clinical ketosis occurring in the first 30 days after calving (Table 2). First, based on parity, each cow has a probability to get SCK. Second, each cow with SCK has, based on parity, an additional probability to get mastitis, metritis, DA, lameness or clinical ketosis (Table 1). Although in practice cows might get more than one related disease after SCK in the first 30 days, no data were available about probabilities of multiple diseases after SCK, nor on the interactions of these multiple diseases on production. Therefore, in the developed simulation model each cow could get only one of these related diseases after SCK. The day of occurrence of SCK was set at day 5 after calving (McArt et al., Reference McArt, Nydam and Oetzel2012). Related diseases only occurred after SCK. Thus, a cow could have no disease, SCK only, or SCK and a related disease. Cows without SCK were excluded from further analysis. Probabilities of SCK and related diseases per parity in the first 30 days after calving were computed using data of Berge and Vertenten (Reference Berge and Vertenten2014). Disease incidence at herd level varies across literature. To represent an average incidence rate while accounting for the difference in risk ratio (RR) across parities and across different combinations of diseases, disease probabilities per parity found by Berge and Vertenten (Reference Berge and Vertenten2014) were up- or downscaled to match average disease incidence at herd level found in literature (Table 2). The additional probability to get a related disease because of SCK was estimated as the difference between the probability to get this disease with SCK and the probability to get it without SCK. It was assumed that only cows with SCK could get clinical ketosis. Third, each cow with SCK and a related disease has a probability to get culled or to die on the farm, together called removal. These probabilities depend on the parity of the cow and the related disease, and were based on removal rates in the first 2 months of the lactation (Table 1). In practice, cows that suffered SCK and a related disease may be removed at a later stage of the lactation, for example, because of related fertility problems. No data, however, were available to relate removals at a later stage to SCK. Therefore, in the simulation model the day of removal was set at day 30 after calving.

Table 2

Disease incidence and consequences on production parameters used to estimate the economic impact of subclinical ketosis and related diseases in dairy cows

¹ Total disease incidence in the first 30 days after calving in the herd.

The discrete events on getting SCK, a related disease after SCK and removal from the herd are simulated by using random numbers drawn from uniform distributions.

Losses due to diseases (part three)

Cows with SCK have a reduced milk production (%/day), which is disease specific (Table 1). During the first 30 days of the lactation, each cow with SCK has the same reduced milk production, irrespective of whether the cow has SCK only or SCK and a related disease, because no data were available to differentiate in milk losses in the first 30 days. Cows with SCK and a related disease also have a reduced milk production after day 30 for a period which is disease specific. These cows receive a disease specific treatment. This treatment is based on the average treatment protocols provided by seven veterinarians, who were randomly selected from different clinics in the Netherlands, and on information from literature. If the cow is treated with antibiotics, the milk is discarded for the prescribed days (Table 2). When a cow is removed, she is immediately replaced by a heifer with average production parameters. A new heifer, however, has a lower milk production than a cow from parity 2 to 5+ and, therefore, removal results in milk losses. Cows with SCK that are not removed from the herd have a reduced fertility. Reduced fertility is included in the model by a prolonged CI of the cow (Table 2). It was assumed that the number of inseminations did not change. The dry period of a cow with SCK did not change and was set to 60 days, but the duration of the lactation period was increased. The prolonged CI depends on the (combination of) disease(s). As no data were available on the combined effect of SCK and related diseases on CI, the impact on CI was determined by adding all the extra days of the disease with the largest effect on CI, and half of the extra days of the other disease, which was adapted from the method of Gröhn et al. (Reference Gröhn, Rajala-Schultz, Allore, DeLorenzo, Hertl and Galligan2003).

After summing all the milk losses, milk production and net energy requirement per lactation are determined and converted to an average per year based on the CI of the cow. Milk losses per cow per year are estimated as the difference between the milk production of the cow with and without SCK.

Calculation of costs (part four)

The total costs of SCK (TCsck _i ) (equation 1) for cow (i) with SCK are estimated as

(1) $${\rm TCsck}_{i} \,{\equals}\,{\rm Cmilk}_{i} {\plus}{\rm Ctreat}_{i} {\plus}{\rm Cmilkmd}_{i} {\plus}{\rm Crem}_{i} {\plus}{\rm Cci}_{i} $$

where Cmilk are costs from reduced milk production (equation 2), Ctreat treatment costs, Cmilkmd costs related to discarded milk (equation 3), Crem costs from removal and Cci costs related to a prolonged CI (equation 5).

Costs for reduced milk production are estimated as:

(2) $${\rm Cmilk}_{i} \,{\equals}\,{\rm ML}_{i} {\times}{\rm MP}{\minus}{\rm FIDL}_{i} {\times}{\rm FP}$$

where ML _i is the reduced milk production in kilogram of cow (i) with SCK, MP the milk price per kilogram of milk, FIDL _i the reduced feed intake in kilogram DM because of ML of a cow (i) with SCK and FP the feed price per kilogram DM of the diet.

Milk price was based on average price of milk from the last 9 years (KWIN-V, 2014).

A diet was composed to estimate the costs related to feeding. A diet for the summer and winter period consisted of concentrates and roughage. Roughage consisted of grass, grass silage and maize silage (CBS, 2014). Energy content per kilogram of dry matter feed was estimated for the summer (6.78 MJ/kg DM) and the winter period (6.53 MJ/kg DM) based on a weighted energy content of the feed ingredients (Table 3). Costs of feed were 0.163 €/kg DM in the summer period and 0.195 €/kg DM in the winter period (Table 4). Based on the estimated net energy requirement of the cow and the net energy content of the feed, DM intake per cow per year was estimated for the same cow with and without SCK. It was assumed that the feed intake of the cow was equal to the energy requirement of the cow. Reduced feed intake per cow per year was estimated as the difference between DM intake of the same cow with and without SCK. The energy requirements for maintenance were the same for cows with or without a disease, while the energy requirement for milk changed with a change in milk yield. As a result, cows with SCK were found to have a lower feed efficiency and a higher energy requirement per kg of milk produced.

Table 3

Dietary composition and feed characteristics used to assess the economic impact of subclinical ketosis and related diseases in dairy cows

¹ Vellinga et al. (Reference Vellinga, Blonk, Marinussen, Van Zeist and De Boer2013).

² KWIN-V (2014).

³ CBS (2014).

⁴ Average taken from 4 years (Nevedi, 2015).

Table 4

Monetary values used to assess the economic impact of subclinical ketosis and related diseases in dairy cows

¹ Feed price is based on a weighted average of feed composition.

Treatment costs (Ctreat _i ) are estimated for cows that had a related disease after SCK. The costs for treatment were based on costs of medication, extra labour for the farmer and the consultation of the veterinarian (Table 4). Cows that are treated receive antibiotics and milk is discarded during this period. All discarded milk is considered waste.

Costs for discarded milk are estimated as:

(3) $${\rm Cmilkmd}_{i} \,{\equals}\,{\rm Mmd}_{i} {\times}{\rm MP}$$

where Mmd _i is the kilogram of discarded milk of cow (i).

Costs of removal (Crem) consist of costs because of milk losses (Cmilkrem) (equation 4) and extra removal costs, because of early introduction of a new heifer.

Costs for milk losses because of removal are estimated as:

(4) $${\rm Cmilkrem}_{i} \,{\equals}\,{\rm MLrem}_{i} {\times}{\rm MP}{\minus}({\rm FIDrem}_{i} {\times}{\rm FP})$$

where MLrem _i is the milk losses in kilogram of cow (i) because of removal, FIDrem _i the reduced feed intake in kilogram DM because of MLrem of cow (i) that was removed.

Cows that are removed before the end of parity 5 are assumed to be removed too early, resulting in extra costs for raising extra heifers. Extra removal costs are estimated with a depreciation method (Getaneh et al., Reference Getaneh, Mekonnen and Hogeveen2017). First, the difference between the opportunity costs of a new heifer and the slaughter price of a heifer is estimated. Second, this difference is depreciated over the total amount of milk that is produced by the cow at the moment of removal. For cows that are removed at parity 2+, average milk production levels were assumed for all previous parities.

The opportunity costs of a new heifer is based on the average price of a heifer from the period 2005 to 2014 (LEI, 2014). The slaughter price of a heifer consists of the slaughter weight of the heifer multiplied with the meat price per kilogram slaughter weight. Slaughter weight of the heifer is estimated with an assumed dressing percentage of 60% (Rutten et al., Reference Rutten, Steeneveld, Inchaisri and Hogeveen2014). It is assumed that the meat price per kilogram slaughter weight of a cow with a disease (€1.82/kg slaughter weight) is lower than of a cow without a disease. This revenue foregone is added when the cow is culled and when the cow dies, the slaughter price of the cow is added. Costs for disposal of the dead cow are also included (Rendac B.V., 2015).

Costs for a prolonged CI (Cci _i ) consist of milk losses because of a prolonged CI and the difference in annual number of calves (equation 5). Reduced milk production that occurred in the extended period of the CI is allocated to Cci.

(5) $${\rm Cci}_{i} \,{\equals}\,{\rm MLci}_{i} {\times}{\rm MP}{\minus}{\rm FIDci}_{i} {\times}{\rm FP}{\plus}{\rm Calfd}_{i} {\times}{\rm CalfP}$$

where MLci _i is the milk losses in kilogram of cow (i) because of prolonged CI, FIDci _i the reduced feed intake in kilogram DM because of MLc _i of cow (i), Calfd _i the reduced number of calves and CalfP the price per calf.

Total costs of SCK are expressed per case of SCK per year and will be presented in the results as an average with the 5% to 95% percentile between brackets. In addition, the median is given in tables. Costs of SCK were analysed per parity, per different costs contributor and per disease (combination).

Sensitivity analysis

A sensitivity analysis was performed to analyse the impact of changes in input parameters on the economic impact of SCK. Input parameters of the reference situation (Tables 1 to 4) were adjusted to alternative values and the impact of alternative values were analysed separately.

Probabilities of each disease and probabilities of removal of all diseases were adjusted. Probabilities of getting mastitis, or metritis, or DA, or lameness or clinical ketosis were increased and decreased with 50% (increased P per disease), whereas probabilities of removal were doubled (increased P removal).

The RR and removal probabilities of metritis and DA were increased based on values found in literature. The RR of metritis for cows with SCK was increased from 1.4 to 3.3 (increased risk ratio metritis) (Duffield et al., Reference Duffield, Lissemore, McBride and Leslie2009). The RR of DA for cows with SCK was increased from 3.7 to 8.0 (increased risk ratio DA) (LeBlanc et al., Reference LeBlanc, Leslie and Duffield2005). Furthermore, in each parity removal risk of DA and metritis were increased to 23.4% (DA) and to 6.0% (metritis) (increased P removal metritis and DA) (Raizman et al., Reference Raizman, Santos and Thurmond2002; Overton and Fetrow, Reference Overton and Fetrow2008).

Milk price and opportunity costs of a heifer were adjusted. Milk price was adjusted from 0.344€/kg milk to the highest (0.416 €/kg milk) (increased milk price) and lowest (0.269 €/kg milk) (decreased milk price) milk price from the last 9 years in the Netherlands (KWIN-V, 2014). Opportunity costs of the heifer were increased from €950 to €1790 (increased opportunity costs heifer) (Mohd Nor et al., Reference Mohd Nor, Steeneveld, Derkman, Verbruggen, Evers, de Haan and Hogeveen2015).