Show simple item record

dc.contributor.advisorMorley, Paul
dc.contributor.authorPoppy, Gerald
dc.date.accessioned2017-06-09T15:41:13Z
dc.date.available2019-06-06T22:59:22Z
dc.date.submitted2017
dc.identifierPoppy_colostate_0053A_14100.pdf
dc.identifier.urihttp://hdl.handle.net/10217/181373
dc.description2017 Spring
dc.descriptionIncludes bibliographical references.
dc.description.abstractAnimal agriculture in today's economic environment is often complex and the uncertainties involved in the decision process make being profitable a challenge. Serving as business consultants, veterinarians can aid producers in helping to make profitable decisions by utilizing available decision tools that enable a better understanding of the economic risk for decisions. Scientific studies that examine the biological response to health or management interventions on dairy farms, while valuable for understanding biology are sometimes limited in their ability to aid in the making good decisions for interventions in agriculture. Adding economics as well as incorporating the variance associated with point effect estimates of biological effect may be a way to decrease the uncertainty or better understand the risk surrounding a management decision. One decision tool available for understanding possible interventions is the use of cross sectional surveys and longitudinal observational studies. A longitudinal study was designed to evaluate various management factors and feed additives and their association with undifferentiated diarrhea events on dairy farms. Based on data from 76 farms, our research team found that a fermented Saccharomyces cerevisiae yeast culture (SCFP) reduced the risk of a cow having a diarrhea event by 30% (IR = 0.707 (P = 0.043, CI = 0.505, 0.989). In addition, having a herd located in the Eastern US versus the Western US was associated with more diarrhea events (IR= 2.036 P = 0.066, CI = 0.953, 4.39). In striving to find the best literature and studies available to help guide the decision process, published studies may differ in estimates of the magnitude of herd response to various management inputs (actions). One key tool that is gaining scientific prominence is the use of meta-analytic techniques to combine multiple studies into a single entity to predict the effect of certain interventions on certain indices of herd health and productivity. A meta-analysis of thirty-six separate studies on a Saccharomyces cerevisiae yeast culture fermentation product was conducted. A total of 69 comparisons met the criteria for inclusion in a random-effects meta-analysis and a sub-group analysis of peer reviewed studies of feeding a SCFP showed an estimated raw mean difference between treated and untreated cattle of 1.18 kg/d (95% CI, 0.55 to 1.81), 1.61 kg/d (95% CI, 0.92 to 2.29), and 1.65 kg/d (95% CI, 0.97 to 2.34) for milk yield, 3.5% fat corrected milk and energy corrected milk, respectively. Milk fat yield and milk protein yield showed an increase in the raw mean difference of 0.06 kg/d (95% CI, 0.01 to 0.10) and 0.03 kg/d (95% CI, 0.00 to 0.05). Estimated raw mean difference in dry matter intake during early lactation (< 70 DIM) and non-early lactation were 0.62 kg/d (95% CI, 0.21 to 1.02) and a decrease of 0.78 kg/d (95% CI, -1.36 to -0.21), respectively from feeding SCFP. Another meta-analysis of active dry yeast (ADY) products was performed; this included 22 papers with 25 comparisons that met the final criteria for inclusion. These studies, conducted in 13 different countries, evaluated active dry yeast products from 7 different companies. This random-effects meta-analysis, showed there was high heterogeneity in the study outcome for milk yield, making it an unreliable outcome to report. One sub-group analysis identified an area of heterogeneity to be study location (in North America versus outside North America). Milk yield for the 7 studies conducted in North American were 0.49 kg/d versus 0.96 kg/d for 13 studies conducted outside North America. The raw mean difference in milk fat yield was 0.05 kg/d and there was a numerical difference in milk protein yield of 0.02 kg/d. No difference in dry matter intake was observed. Utilizing the information in meta-analysis of products can be improved by the use of stochastic analysis by incorporating the variance from the point estimate parameters into a partial budget of the production changes. Software programs exist that can perform Monte Carlo simulations on partial budgets, factoring in both the biological effects and their variance from the meta-analysis result as well as the economics of the biological change for the producer's business. ModelRisk 5.1.1 (Vose Software BVBA, Belgium, 2015) was used to generate 10,000 iterations of a partial budget, utilizing the mean outcome and variance parameters from the SCFC meta-analysis. The resulting stochastic partial budget calculation showed a risk of not having above a break-even response as 0.27%; in addition, the cost of making a Type 1 error versus a Type 2 error would be less than $0.001 versus $0.38 per cow/d. This means that based on the information contained in the meta-analysis the producer is left with the probability of 0.27% of losing <$0.001 / cow /d by feeding SCFP versus the decision to not feed SCFO and have a 99.8% chance of not earning $0.38 per cow/d. Based on the meta-analysis data, a Monte Carlo simulation of the ADY products in early lactation showed a risk of not having a break-even response as 38.87%. This dissertation demonstrates the use of direct fed microbials may have a benefit in nutrition programs on dairies. Specifically, the use of SCFP was associated with a decrease in diarrhea events as well as increases in milk production when analyzed using meta-analytic methodology. To aid in decision making the use of stochastic analysis utilizing the variance from the meta-analysis along with the associated point effects is a useful tool to graphically and numerically demonstrate the uncertainty of the outcome. Integrating the biological variation and it associated economic values into a distribution of outcome along with their associated conditional probabilities can be used to calculate the cost of Type 1 and Type 2 components of the decision helping to frame the decision in quantifiable units possibly more useful to a diary producer.
dc.languageEnglish
dc.language.isoeng
dc.publisherColorado State University. Libraries
dc.rightsCopyright of the original work is retained by the author.
dc.subjectDecision Management
dc.subjectMeta-Analysis
dc.subjectYeast Culture
dc.subjectDiarrhea
dc.subjectDairy Management
dc.subjectStochastic Analysis
dc.titleDairy management decisions utilizing available evidence and information
dc.typeThesis
dc.identifier.schemaETD Data Dictionary 1.1
dc.rights.accessEmbargo Expires: 06/06/2019
dcterms.embargo.terms2019-06-06
dcterms.embargo.expires2019-06-06
dc.contributor.committeememberHill, Ashley
dc.contributor.committeememberVan Metre, Dave
dc.contributor.committeememberPendell, Dustin
dc.contributor.committeememberGroenendaal, Huybert
thesis.degree.nameDoctor of Philosophy (Ph.D.)
thesis.degree.levelDoctoral
thesis.degree.disciplineClinical Sciences
thesis.degree.grantorColorado State University


Files in this item

Thumbnail

This item appears in the following Collection(s)

Show simple item record