Variation in partial direct costs of treating clinical mastitis among 37 Wisconsin dairy farms

The objective of this observational study was to describe variations in partial direct costs of clinical mastitis (CM) treatments among 37 dairy herds using data obtained from herd management records. Animal health and drug purchase records were retrospectively collected from 37 Wisconsin dairy herds for a period of 1 yr. Each farm was visited to verify case definitions, recording accuracy, and detection criteria of CM cases. Descriptive statistics were used to summarize cost of drugs and milk discard. Differences in costs among protocols, intramammary (IMM) products, parities, days in milk, and recurrence were analyzed using ANOVA. Of 20,625 cases of CM, 31% did not receive antimicrobial treatment. The average cost of drugs and milk discard (including cases that were not treated) was $192.36 ± 8.90 (mean ± SE) per case and ranged among farms from $118.13 to $337.25. For CM cases treated only with IMM antimicrobials, milk discard accounted for 87% of total costs and was highly influenced by duration of therapy. Differences in costs were observed among parities, recurrence, and stage of lactation at case detection. Eight different treatment protocols were observed, but 64% of cases were treated using only IMM antimicrobials. Treatment costs varied among protocols; however, cases treated using both IMM and injectable antimicrobials as well as supportive therapy had the greatest costs as they were also treated for the longest duration. Ceftiofur was used for 82% of cases that received IMM antimicrobials while ampicillin was used for 51% of cases treated using injectable antimicrobials. With the exception of ceftiofur and pirlimycin IMM products, many IMM products were given for durations that exceeded the maximum labeled duration. For cases treated using only IMM therapy, as compared with observed costs, we estimated that partial direct costs could be reduced by $65.20 per case if the minimum labeled durations were used. Overall, partial direct costs per case varied among herds, cow factors, and treatment protocols and were highly influenced by the duration


INTRODUCTION
Dairy farm productivity and animal welfare are impaired by the occurrence of disease and avoiding unnecessary costs associated with treatment is essential to maintain profitability.Clinical mastitis (CM) occurs in about 25% of lactating cows each year (USDA et al., 2018;Gonçalves et al., 2022) and is detected on almost every dairy farm.Mastitis is an expensive disease, and costs of CM include direct costs, which include milk discard, drugs, diagnostic testing, and labor, as well as indirect costs, such as reduced milk production, reduced value of milk, increased risk of culling (Pinzón-Sánchez et al., 2011), and reduced fertility (Fuenzalida et al., 2015).Indirect costs have been estimated to account for >70% of total costs of CM (Rollin et al., 2015) but direct costs are most obvious to farmers.
The overall economic impact of CM has been evaluated (Halasa et al., 2007;Morse et al., 1987;Aghamohammadi et al., 2018) and total losses due to CM are often estimated by developing models using assumptions obtained from scientific literature.In these studies, estimated total costs of CM have ranged from approximately $40 to $500 per case depending on the type of model, variables used to estimate costs, types of treatment decisions, and the duration of treatment (Pinzón-Sánchez et al., 2011;Rollin et al., 2015;Liang et al., 2017).Of variables used to estimate costs of CM, expenses related to purchase of drugs, milk discard, labor, diagnosis, and culling are often included (Seegers et al., 2003;Huijps et al., 2008).Costs associated with milk discard are an important determinant of CM costs.Abnormal milk may not be sold for human consumption, and milk from cows receiving antibiotics must be discarded for prescribed withdrawal periods.Thus, increased monetary losses are associated with longer durations of treatment (Pinzón-Sánchez et al., 2011).For example, expected monetary losses of CM were estimated to increase from approximately $65 to $187 per case as treatment days increased from 2 to 8 d (Pinzón-Sánchez et al., 2011).
Antimicrobials are usually used for treatment of CM and account for about 35% of total antimicrobial doses used on dairy farms (Pol and Ruegg, 2007;Stevens et al., 2016;Leite de Campos et al., 2021).Consumers have expressed considerable concern about usage of antimicrobials on dairy farms (Wemette et al., 2021) and reductions in antimicrobial usage must address use of antibiotics for treatment and prevention of mastitis.Not all cases of CM require antimicrobial therapy to achieve successful outcomes (Lago et al., 2011a,b;Fuenzalida and Ruegg, 2019;Ruegg, 2021).Reducing unnecessary usage of antimicrobials by identifying cows that will not benefit from antimicrobial treatment and reducing treatment duration (when feasible) will reduce costs of treatment as well as the number of antimicrobial doses that are used on dairy farms (Ruegg, 2021).A better understanding of partial direct costs associated with treatment of CM will help dairy farmers and their advisers make informed decisions about the economic impact of longer duration therapy.The objective of this retrospective observational study was to describe drug and milk discard costs associated with treatment of CM on 37 dairy herds using data obtained from herd management records and drug purchase receipts and compare variation in components of treatment cost and treatment protocols among herds.

Herd Selection and Data Collection
Partial direct costs of CM treatment were defined as the costs associated with milk discard (due to abnormal milk and the milk withholding period when drugs are administered) and the purchase price of drugs.Additional costs such as labor and diagnostics were not included in this study as these components are often standardized and have little variation among herds.These costs were estimated using computerized herd records obtained from 37 large dairy herds in Wisconsin.Herd selection and data collection have been previously described (Leite de Campos et al., 2021) and data collection was approved by the Institutional Review Board at the University of Wisconsin-Madison (Madison, WI;2017-1333-CR002)-CR002).In brief, 40 dairy herds were recruited based on herd size (≥250 lactating cows) and availability of drug purchase, computerized animal health, and treatment records.Treatment records were retrospectively collected for a period of 1 yr and a single farm visit was performed to verify treatment records and case definitions.During our farm visit, animal health managers were surveyed about herd characteristics, detection procedures, and treatment protocols.Before the farm visit, backup files containing electronic dairy health records were sent to researchers who reviewed and printed all potential animal health events, which were then discussed during a 2 h interview for which each farmer received $100 honorarium.For this study, 37 of the original 40 herds were selected based on usage of a common dairy management program (Dairy Comp 305; Valley Agricultural Software, Tulare, CA).Data on each case of CM and monthly DHIA milk production per cow were obtained from herd records.Data entered into herd records were verified during the herd visit and when electronic records did not list specific antimicrobials used or the duration of treatment, that information was imputed based on responses obtained during the interview.Of 37 herds, electronic records for 21 included the duration of intramammary (IMM) antimicrobial treatment for almost all (>96%) CM events, electronic records for 2 herds included duration of IMM antimicrobial treatment for the majority (>57%) of CM events, while electronic records for 14 herds did not include information on duration of treatment.Overall, duration of IMM antimicrobial treatments was downloaded from electronic records for 59% of CM events and duration of treatment for the remaining 41% of IMM treatments were imputed by using the average number of treatment days described by interviewing animal health managers on each farm.
Partial direct treatment costs for CM were calculated for all cases, including those treated using antimicrobials, as well as cases that received supportive therapy or no treatment.For cases that did not receive antimicrobials, partial direct costs of each case of CM were estimated based on costs of milk discard and supportive treatment (when given).For cases that received antimicrobial treatment, costs of IMM treatment (if any), injectable (INJ) treatment (if any), supportive treatment (if any), and milk discard were used.
Each case of CM was defined at cow-level and included all treatments that occurred within 14 d from the initial date that the case was recorded.When multiple treatments were given within the 14 d period, milk discard was determined based on the beginning of the first treatment to longest withholding period for any drug given.Recurrent cases of CM were defined when a case occurred in the same cow >14 d after the first CM event.

Cost of Drugs
Costs of drugs used for treatment of mastitis were estimated using purchase records obtained during the farm visit.To standardize purchase prices among farms, prices were summarized per product and the overall average price was used to calculate the costs across all herds in the study.Costs of IMM treatments were estimated using this formula: IMM Cost = quarters treated × days of treatment × applications per day × price per intramammary antibiotic tube.
Costs of INJ and oral treatments were estimated based on the dose given and calculated using the following formula: INJ Cost = maximum dose × days of treatment × applications per day × price per mL.

Milk Discard
Monthly test-day milk production for each cow was obtained from electronic records by creating items and using the following command in the dairy management software (DairyComp305; Valley Agricultural Software, Tulare, CA): EVENTS\2S365I ID BDAT LACT ITEM1 ITEM2 ITEM3 FOR LACT > 0.
The value of the test-day milk production closest to the event of CM was used to estimate the amount of milk discarded.Values within a range of −14 d before a case of CM to +30 d after the event were used.If a cow had more than one test day during this period, the value from the closest test date after CM was selected.For cases that did not have a value for milk within this range, herd average milk production was used to estimate milk discard.Across all herds, milk discard values for 90% (n = 18,609) of CM cases were within the specified range, while the herd average milk production was used to estimate milk discard for the remaining 10% (n = 2,016) of cases.
For animals treated with any medication that required milk discard (cases that followed or exceeded label direction), total days of milk discard was defined as the sum of treatment days plus withholding days.
Withholding period was determined using the current label listed in the US National Library of Medicine (https: / / www .nih.gov/).For animals that were not given any medication that required milk discard, we estimated a milk discard period of 5 d based on expected period of abnormal milk (Oliveira and Ruegg, 2014).A 5-d milk discard after treatment was used for systemic treatments using florfenicol and spectinomycin, as they do not have a labeled milk withholding period.

Statistical Analysis
Statistical analyses were performed using SAS version 9.4 (SAS Institute Inc., Cary, NC).Descriptive characteristics were summarized using PROC FREQ.The cost of CM treatment per case was summarized for each farm using PROC MEANS.Costs were estimated on a per-case basis and summarized for cases that received solely IMM treatments; solely INJ treatments; solely supportive treatments; combination of IMM and INJ treatments; combination of IMM and supportive therapy; combination of IMM, INJ, and supportive therapy; combination of INJ and supportive therapy; and for cases that did not receive any therapies.
For CM cases that did not receive antimicrobials, costs per case were calculated by summing the total cost of milk discard and supportive treatments (if any) and dividing by the sum of cases that received supportive therapy only and cases that did not receive any antimicrobials.Partial direct costs of treatment of CM cases that received antimicrobials were calculated by summing the cost of milk discard and treatments and dividing by the number of cases that received antimicrobials.At the farm level, overall treatment costs were calculated by summing the milk discard and drug costs of treated and nontreated cases and dividing by the total number of CM cases occurring on each farm.
Herd size and CM rate were classified based on median values and associations between herd-level cost per case and these variables were analyzed using ANOVA.Differences in observed partial direct costs per case were compared with estimated treatment costs if IMM treatments had been used according to minimal approved label duration of treatment.Label instructions for IMM products were obtained from the US National Library of Medicine and a minimum of 2 d duration of treatment was chosen for ceftiofur hydrochloride and pirlimycin hydrochloride products (each have an approved range treatment duration of 2-8 d).Difference in means between observed cost per case and costs of treatment using label durations were analyzed using a paired t-test.
PROC UNIVARIATE was used to assess normality of data and differences in partial direct costs among products and protocols were evaluated using a model that used cost per case as the dependent variable and products or treatment protocols as independent variables.Difference in costs per case based on parity groups (parity 1, 2, ≥ 3), recurrence (first; second; third or greater), and stage of lactation at case detection (0-30 DIM; 31-60 DIM; 61-90 DIM; 91-120 DIM; 121-150 DIM; 151-180 DIM; 211-240 DIM; 241-270 DIM; >270 DIM) were analyzed using PROC GLIMMIX in models that used cost per case as the dependent variable and parity, recurrence, or stage of lactation as independent variables.For all models, farm was included as a random effect and differences among least squares means were adjusted using Tukey method (SAS, 2013).Statistical analyses were performed only for products or treatment protocols used on ≥5 herds.The null hypothesis was that cost per case did not vary among products and statistical difference was considered when P < 0.05.

Herd Description
Descriptive statistics of the original 40 herds have been previously reported (Leite de Campos et al., 2021).Herds (n = 37) included in this study contained an average of 1,351 ± 180 lactating cows per farm (ranging from 332 to 5,005) and were primarily composed of Holstein cows (98% ± 0.76).The average bulk tank SCC among herds was 143,000 cells/mL ± 7,230 (ranging from 77,000 to 320,000 cells/mL) and rolling herd average for milk was 13,377 kg ± 164 (ranging from 10,829 to 15,059 kg).All herds housed lactating cows in freestalls and the primary bedding type was fresh sand (n = 20), followed by recycled sand (n = 9), manure solids (n = 3), wood products (n = 2), or mixed bedding materials (n = 3).Average daily milk produc-tion of affected cows before occurrence of CM was 43 kg ± 0.71 (Table 1).Of 20,625 cases of CM, 6,473 (31%) did not receive antimicrobial treatments while 14,152 (69%) were treated using antimicrobials (Table 1).Except for one farm that recorded very few cases of CM and only used supportive products, all farms used IMM antimicrobials for treatment of at least one case of CM.Injectable antimicrobials were used for treatment of at least one case of CM in 24 herds (65%) and one farm used an oral antimicrobial for treatment of 2 cases of CM.Supportive treatment was used by 15 farms (41% of farms).

Overall Cost Per Case of CM
Across all farms and all treatment protocols, the average partial direct cost of treatment per case per farm was $192.36 ± 8.90 and ranged from $118.13 to $337.25 (Figure 1).For herds (n = 36) that reported administration of antimicrobials (either IMM or INJ) for treatment of CM, the average partial direct costs per treated case were $209 ± 9.65 and ranged from $119.95 to $372.26.For herds (n = 25) that reported CM cases not treated with antimicrobials, the average partial direct costs per nontreated case were $87.44 ± 4.34 and ranged from $7.99 to $119.59 (Table 1).
Overall partial direct costs of treatment per case (including cases treated with antimicrobials and not treated cases) were not associated with herd size or CM rate.The median value for herd size and CM rate across herds was used as a threshold for the analyses.The average costs for herds with ≤954 cows (n = 20) were $186.0 ± 9.74 and $199.80 ± 15.77 for herds with >954 cows (n = 17; P = 0.44).The average costs for herds with a CM rate ≤31% (n = 20) were $179.50 ± 10.99 and $207.50 ± 13.89 for herds with CM rate >31% (n = 17; P = 0.52).

Cow Factors Influencing Cost of Treatment
Cases were distributed among cows in first lactation (n = 3,973 cases; 19%), second lactation (n = 6,711; 33%), and third or greater lactation (n = 9,941, 48%).Treatment costs varied among parities, and cost per case was greater for cows with ≥3 lactations ($196.02 ± 8.56  Clinical mastitis (CM) rate was calculated as the number of cases divided by the number of adult cows (dry and lactating cows).A CM case was defined as all cases that occurred within 14 d from the first event.
2 Average number of days milk was discarded, calculated as the sum of treatment days plus withholding days.
3 Average milk production (kg) of CM cases obtained from monthly DHIA in period of 14 d before CM to 30 d after. 4 Proportion of CM cases treated with injectable antimicrobials.

5
Proportion of CM cases treated with any antimicrobial.

6
Proportion of CM cases treated with supportive therapy.

7
Average cost per case for CM events that did not receive antimicrobial treatment.Average cost per case was calculated by dividing the total cost with milk discard for a period of 5 d and nonantibiotic supportive treatments (if any) by the frequency of cases that did not receive antimicrobials.
8 Average cost per case for CM events treated with antimicrobials (intramammary; injectable; or intramammary + injectable).Average cost per case was calculated by dividing the total cost with milk discard and any treatment (intramammary or injectable antimicrobials; supportive products) by the frequency of treated cases.9 Average cost per case for all CM cases.Average cost per case was calculated by dividing the total cost of cases treated and not treated with antimicrobials by the frequency of all CM cases on farm.
The average DIM at occurrence of CM was 162 ± 0.77 and average milk production during this period was 41 kg ± 0.09.Days in milk at occurrence of a case ranged from 0 to 948, and 25% of CM cases occurred within 78 d after calving, 50% of CM cases occurred within 147 d after calving, and 75% of the CM cases occurred within 232 d after calving.Treatment costs per case varied across DIM, and average cost per treatment was greater at 91 to 120 DIM ($203.49± 9.21) than when DIM was >270 ($151.32 ± 9.24; P < 0.001; Figure 2).

Comparisons Among Treatment Protocols
The most common treatment protocol was administration of only IMM antimicrobials which was reported by 35 farms and accounted for 64% of cases (Table 2).The average days of discard (duration of treatment plus the withholding period) was 9.3 ± 0.35 d, and costs per case were $184.93 ± 7.76, ranging from $126.97 to $320.12 (Table 2).For these cases, costs were distributed as 13% for IMM antimicrobials ($23.91 ± 1.58) and 87% for milk discard ($161.50 ± 6.54).
Cases that did not receive any antimicrobial or supportive therapies were reported by 22 farms and accounted for 30% of all CM cases.The average costs per nontreated case were $79.01 ± 9.47 ranging from $61.73 to $110.54 (Table 2).The entire partial direct cost of these cases was based on milk discard.
Combined treatments using both IMM and INJ antimicrobials were reported by 22 farms and accounted for 3% of all CM cases.The average days of discard was 9.1 ± 0.58 d and costs per case were $206.70 ± 9.47, ranging from $96.54 to $400.85 (Table 2).In this scenario, costs were distributed as 31% for IMM and INJ antimicrobials ($63.37 ± 9.16) and 69% for milk discard ($139.64 ± 9.78).
Cases treated only with supportive products were reported by 12 farms and accounted for 1% of all CM cases.The average days of discard was 4.9 ± 0.23 d and cost per case were $82.89 ± 12.41, ranging from $7.99 to $130.39 (Table 2).In this scenario, costs of supportive products were 14% ($13.66 ± 4.43), while costs of milk discard accounted for 86% of total costs ($82.56 ± 8.72).
Cases treated only using INJ antimicrobials were reported by 9 farms and accounted for less than 1% of all CM cases.The average days of discard was 5.9 ± 0.43 d and costs per case were $130.25 ± 14.17, ranging from $82.63 to $193.09 (Table 2).In this scenario, costs of INJ antimicrobials accounted for 21% of the total ($26.78 ± 11.11), and the costs of milk discard accounted for 79% of the total ($103.71 ± 11.30).
Treatment using IMM and INJ antimicrobials, in combination with supportive therapies were reported by 3 farms and accounted for less than 1% of all CM cases.The average days of discard was 11.5 ± 2.48 d and costs per case were $240.03± 30.73, ranging from $207.75 to $301.47 (Table 2).In this scenario, costs were distributed as 47% for antimicrobial and supportive products ($43.31 ± 10.38 for IMM antimicrobials, $56.04 ± 29.04 for INJ antimicrobials, $13.18 ± 4.66 for supportive products) and 53% for milk discard ($127.50 ± 12.86).
Only 15 CM cases from one herd were treated solely with INJ antimicrobials and supportive products.The average days of discard was 6.6 d and costs per case were $162.67.In this scenario, costs of INJ antimicrobials accounted for 16% of the total ($26), supportive products accounted for 9% of the total cost ($14.04),and costs of milk discard accounted for 75% of the total ($122.63;Table 2).

IMM Products and Estimated Cost of IMM Treatment When Following Label Directives
Among herds that used IMM antimicrobials for treatment of CM (n = 36 herds), 6 different IMM products were reported (Table 3).Ceftiofur was the most commonly used antimicrobial, and was administered in 32 herds, representing 82% of the IMM treatments (Table 2).Other IMM products included, cephapirin (used in 13 herds for 9% of treatments), hetacillin (9 herds, 3% of the treatments), pirlimycin (12 herds, 4% of the treatments), amoxicillin (6 herds, 1% of the treatments), and cloxacillin (1 herd, <1% of the treatments).Based on observed costs for CM cases treated with IMM antimicrobials only, no difference in costs were observed by active ingredient (P = 0.11; Table 3).

Products Used for Injectable or Supportive Treatment
Among herds that used INJ antimicrobial products either solely or combined with other products (n = 24 herds), 8 different antimicrobials were used.Ampicillin was used by 50% (n = 12 herds) and represented 51% of the treatments.Injectable ceftiofur products and oxytetracycline were used by 10 herds each and represented 44% of INJ treatments, whereas florfenicol, spectinomycin, and sulfadimethoxine were each used by 1 herd and together represented approximately 5% of the treatments.
A total of 10 products were included in electronic records as supportive therapy of CM.Among products, hypertonic saline (n = 8 herds; 317 cases) and flunixin meglumine (n = 7 herds; 122 cases) were most commonly used.Additional supportive products include dexamethasone (n = 4 herds; 26 cases), aspirin (n = 2 herds; 15 cases), and others such as oxytocin, vitamin B, and oral calcium supplements that were used on as treatments for 64 cases on <5 herds each.

DISCUSSION
Bulk tank SCC has decreased in US dairy herds indicating control of subclinical mastitis and continued improvements in milk quality (APHIS, 2019).In contrast, the proportion of cows affected with CM appears to be stable or increasing (Gonçalves et al., 2022), probably as a result of exposure to environmental pathogens that tend to cause larger inflammatory responses resulting in nonsevere CM.Mastitis is an expensive disease as IMM infection reduces productivity and results in discarded milk and indirect costs that reduces farm income (Rollin et al., 2015;Aghamohammadi et al., 2018).Maintaining control of costs associated with disease treatment is crucial to maximize farm productivity.Reducing antimicrobial usage to that necessary to maintain animal health is an important aspect of antimicrobial stewardship and demonstration of economic advantages based on reductions in duration or usage of antimicrobials may motivate examination of treatment protocols for mastitis.By calculating the average cost per case based on differences in treatment protocols and among herds, we were able to estimate difference in cost among protocols that may inform future treatment decision and contribute to judicious usage of antimicrobials.
Similar to previous studies, great variation in cost of treatment was observed among herds (Rodrigues et al., 2005;Aghamohammadi et al., 2018).Herds with the greatest costs spent almost 3 times more per case compared with herds with the least costs, and this difference is associated with differences among treatment protocols, including differences in duration of therapy and the proportion of cases that are not treated using antimicrobials.Our cost estimates are partial costs and do not include diagnostic costs used to guide selective therapy programs such as use of culture or other diagnostic tests nor do they include labor.However, based on bulk tank SCC and other characteristics of these herds, these farms likely had minimal prevalence of mastitis caused by pathogens with very low expected spontaneous cure rates (such as Staphylococcus aureus) and thus may benefit from selective use of antimicrobials or shorter duration therapies that result in less discarded milk.
Treatment costs for CM have varied among studies and are highly dependent on the variables and assumptions included in analyses.Direct costs often include costs such as diagnostics, therapeutics, milk discard, veterinary service, and labor.Cost components such as labor and cost of diagnosis are important in determining total cost of CM, but on a practical basis, when comparing costs, those components are often standardized and vary little among cases (Pinzón-Sánchez et al., 2011).As our objective was to determine variation among farms using actual data included in herd records, we chose not to include those costs in our calculation.Variation in direct costs of treatment can be attributed to differences in duration of treatment, milk yield of affected cows, milk price, severity, and choice of therapies.Although most of these variables are not under the control of the farmer, duration of treatment can be controlled.Expected monetary losses of a case of CM have previously been shown to be strongly associated with number of treatment days (Pinzón-Sánchez et al., 2011).Of 6 IMM products used on these farms, 4 have labeled durations of treatment of <3 d but the observed duration of treatment for all these products exceeded label recommendations.Two products have flexible labels with recommended durations of 2 to 8 d but were administered for about 5 d each.Interestingly, one of those products (SpectramastLC, Zoetis) is approved in Canada with a label that prescribes only a 2-d treatment (https: / / www .drugs.com/vet/ spectramast -lc -can .html).For most etiological agents, there is very limited evidence to support improvement in clinical outcomes based on longer duration therapy (Ruegg, 2021), and use of shorter duration therapy (when appropriate) is an obvious way to reduce treatment costs.Others have demonstrated that reduced milk production and cost of  Least squares means values within the same column with different superscripts differ from each other (P < 0.01). 1 Treatment practices that occurred in ≤5 farms were not included in the analysis among means.Milk withhold days is the sum of the number of days actually treated plus the label directed withholding period (for example, amoxicillin was treated for 4.5 d and withheld for 2.5 d).
discarding milk often represents the largest proportion of total cost of treatment (Seegers et al., 2003;Halasa et al., 2007).Two critical elements in estimating costs are the milk price and the average milk produced by affected cows.When milk price is higher, or when more milk is discarded, losses associated with treatment increase.In previous studies that calculated direct treatment cost (drug, milk discard, labor) of CM, cows produced approximately 30 to 37 kg of milk per day, and milk discard represented about 70% of the total cost of CM (Rodrigues et al., 2005;Pinzón-Sánchez et al., 2011).In our study, the treatment protocol most used was administration of IMM products only, and milk discard for these cases represented about 90% of the total cost of treatment.Costs of treating mastitis were greatest during periods of highest milk production and each day of treatment beyond label recommended durations results in additional cost with questionable effects on efficacy.Our results demonstrated that the costs of treatment for cows producing around 25 kg of milk/d would be approximately $62, while for cows producing about 50 kg of milk per day, the cost per treatment would be $174.To make informed treatment decisions, it is important for veterinarians and farmers to understand the impact of extended durations on costs of therapy.
As compared with a national level report (USDA et al., 2018) and Oliveira and Ruegg, (2014), the proportion of cases of CM treated with antimicrobials was less than previously observed.Most cases were treated only with IMM antimicrobials (Oliveira and Ruegg, 2014;Aghamohammadi et al., 2018), and we may have underestimated the proportion of nonantimicrobial treated cases as 14 herds recorded 100% of their cases as treated and may not have recorded CM cases that did not receive treatments.In the United States, only IMM antimicrobials are approved for treatment of mastitis, although extra-label usage of some systemically administered drugs is allowed under veterinary supervision (Ruegg, 2021).Although most treatments were compliant with extra-label guidelines, a few cases received extra-label usage of sulfonamides, which is not permitted under FDA regulations.About 60% of the herds enrolled in our study reported usage of INJ antimicrobials, but only 4% of cases were treated using this route, and it is likely that these cases were more severe.Very few farms recorded severity in the electronic records, and we were unable to determine severity of most cases based on treatment protocol.Use of INJ antimicrobials for treatment of CM are often associated with severe cases (Oliveira and Ruegg, 2014), and although our data do not include severity score, it is likely that INJ treatments are linked with severe cases.Among IMM and INJ products reported in this study, ceftiofur was the primary IMM antimicrobial and ampicillin was the primary INJ antimicrobial (Oliveira and Ruegg, 2014).Clinical mastitis cases had similar characteristics as previously described (Oliveira et al., 2013) and about half of the cases occurred during early to middle lactation.Older cows (parity ≥3) accounted for most of the cases, and as expected, based on greater milk yield in older cows, costs were approximately $40 and $30 greater for cows in parity ≥3 as compared with cows in parity 1 and 2, respectively.Surprisingly, we also observed a reduction in cost per case based on recurrence of CM, and cases that recurred for a second and third time or more were approximately $20 and $42 cheaper than the cost of first case.Similar to difference in cost among parities, difference in cost based on recurrence could be explained by difference in milk production across DIM and milk production reductions caused by CM.
Treatment decisions for CM should be performed based on etiology and ideally will consider cow factors such as stage of lactation, parity, and history of previous clinical or subclinical mastitis (Ruegg, 2021).While duration of treatment is sometimes altered based on pathogen, there is little evidence in support of a meaningful difference of efficacy among different IMM products or for most etiologies evidence to demonstrate benefit from extended duration of treatment (Kolar et al., 2021;Ruegg, 2021).We used the treatment duration indicated on product labels to estimate differences in costs from that we observed.Treatments with IMM cephapirin had the greatest difference between label duration and usage observed in our study, and this result agrees with results from Oliveira and Ruegg (2014), that shows that label compliance for duration of treatment was only followed in approximately 30% of mild and 6% of moderate CM cases treated with cephapirin.

CONCLUSIONS
Our results demonstrate variation in partial direct cost of treatment of CM among larger Wisconsin dairy herds as well as differences in the proportion of cases treated with antimicrobials.Great variation in cost was associated with parity, recurrence of CM cases, DIM at occurrence of the case, and treatment protocol.Milk discard represented the greatest proportion of cost and CM cases treated for longer duration were associated with greater costs.Reducing the duration of treatment to the approved label directives for routine treatments could results in both reduced antimicrobial usage and less losses associated with treatment of CM.

Figure 1 .
Figure 1.Average cost (US$) per case for clinical mastitis cases treated, but not with antimicrobials.Clinical mastitis cases were obtained from 37 Wisconsin dairy farms from September 2017 to December 2017.
Figure 2. Difference in cost (US$) with milk discard and medication for treatment of clinical mastitis by milk yield (kg) across DIM.Least squares means values within DIM with different letters (a-e) differ from each other (P < 0.001).
Leite de Campos et al.: VARIATION IN COSTS OF TREATING MASTITIS Leite de Campos et al.: VARIATION IN COSTS OF TREATING MASTITIS

Table 1 .
Leite de Campos et al.: VARIATION IN COSTS OF TREATING MASTITIS Leite de Campos et al.: VARIATION IN COSTS OF TREATING MASTITIS Characterization of clinical mastitis treatment on 37 dairy herds in Wisconsin; data were collected from September to December 2017 and included all cases of clinical mastitis recorded in computerized records Farm

Table 2 .
Characterization of the treatment practices of cases of clinical mastitis (n = 20,625) obtained from 37 dairy herds in Wisconsin from September 2017 to December 2017

Table 3 .
Cost of treatment (US$) for first cases of clinical mastitis treated only with intramammary products (n = 8,630) based on label directives of approved intramammary antimicrobials for clinical mastitis in the United States and estimated cost of treatment based on data records obtained from computerized records from 36 Wisconsin dairy farms 2