Confidence intervals for the ratio of two Poisson rates under one-way differential misclassification using double sampling

Kahle, David; Young, Phil D.; Greer, Brandi A.; Young, Dean M.

doi:10.1016/j.csda.2015.09.013

Cited by 2 publications

(1 citation statement)

References 18 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…It is possible that there may be a lesser degree of misclassification when owner-reported body condition scores are converted to a binary dependent variable—overweight/ obese vs. all other cats. Double sampling is a method that combines a small validation sample with an error-prone main-study sample, potentially yielding results that are more accurate (45). Rowe et al (40) employed this method in their study on early-life risk factors for feline obesity; 144 cats (of 375, 38%) had both an owner-reported BCS and a vet-reported BCS.…”

Section: Discussionmentioning

confidence: 99%

Owner and Cat-Related Risk Factors for Feline Overweight or Obesity

2019

View full text Add to dashboard Cite

Feline obesity is a highly prevalent disease that poses an urgent and serious challenge. Attempted treatment by weight reduction is often unsuccessful; a new preventative approach that focuses on the role of the owner may be helpful. This study used data collected from an international survey of cat owners designed to assess owner personality and self-control, owner-pet attachment, feeding practices, and the cat's body condition. Owner-reported body condition scores (BCS) of cats were assessed using images adapted from a 5-point BCS system and categorized as a binary dependent variable: overweight/obese (BCS 4–5) and not overweight (BCS 1–3). Owner-reported BCS scores using a verbal BCS scale were also used as a binary dependent variable. Of the 6,835 respondents, 30.5% described their cat as overweight/obese using the visual BCS scale, and 32.5% using the verbal scale. Multivariable logistic regression models were built using stepwise-backward selection. A total of 8 variables were significant using the visual score as the dependent variable, while 11 variables were significant using the verbal score as the dependent variable ( p < 0.05). Low owner conscientiousness was associated with an increased risk of feline overweight/obesity (OR = 1.23, 95% CI 1.10–1.38), whereas preference for delayed reward was associated with a decreased risk (OR = 0.84, 95% CI 0.75–0.96). Contrary to expectation, indulgent (OR = 0.76, 95% CI 0.53–0.91) and inconsistent (OR = 0.86, 95% CI 0.76–0.93) feeding practices appeared protective. Other significant variables ( p < 0.05) included cat-related factors (age, gender, housing, source) and management-related factors (dry diet, supermarket dry diet, raw diet, stealing, hunting, and measuring food with a scoop). A third multivariable analysis was performed, using results from cats classified as overweight/obese using both scoring methods, compared with cats classified as a healthy weight using both scoring methods. A total of 10 variables were found to be significant ( p < 0.05). There was significant overlap of results from all three analyses. The results of this study indicate that feline obesity is a complex problem, with many contributing risk factors. It is essential to recognize the importance of owner characteristics, and that the prevention of obesity in cats may require the development of a range of interventional strategies.

show abstract

Section: Discussionmentioning

confidence: 99%