This survey was designed to answer two main research questions: 1) Is there an association between measurements of feline skull conformation and the occurrence of owner-reported respiratory abnormalities. 2) What is the relationship between brachycephlism and routine management and care provided by the owner? The study design has a number of shortcomings and therefore findings from this study are not considered robust.
In terms of scientific reliability, the study design is not considered adequate to provide reliable answers to the research questions because of issues relating to sample selection, source of data, measurement methods, and statistical analysis methods. The definition for brachycephlism is somewhat arbitrary which would make comparison with other studies difficult.
The unit of analysis is the cat, therefore a representative sample of the cat population for whom the study results are applicable should have been obtained. No data about the basic characteristics of cats e.g. age, and medical history, were provided. Though, it is understood that such data have been collected as part of the survey. Therefore it is not possible to assess the extent to which findings from the study can be generalised.
It is not clear why recruitment was not done through veterinary clinics, where objective outcomes (e.g. Measurement of feline skull) could have been done under controlled conditions, using a standardised protocol to allow for reproducibility of the data and to ensure consistency of measurements. In this study, the researcher relied on photographic images, from the cats' owners, with varying level of quality of the images. This approach is unlikely to yield reliable data.
Data on respiratory abnormalities is considered subjective and bias in response is very likely. The research did not present any additional data to address this issue, for example a summary of characteristics of cats by the cat's owner status (responders vs non responders).
Survey data should be analysed using analysis methods that are design based. The statistical methods used to analyse data from this survey are general. There is no information about the sampling frame work used in this survey and therefore it is not possible to ascertain whether the analysis methods are appropriate.
Note also that variable selections for the multivariate model used 2% significance threshold for p-values in the univariate analysis. The approach to first proceed with a univariate analysis and to use of AIC to find the subset of variables that produce an optimal model is in principle acceptable. However, the use of a more liberal p-value (10% or 20%) for variable screening is recommended since lower p-value can fail to detect variables that are known to be important. Furthermore outcome variables that have an ordinal scale should use methods that preserve the scaling of the variable in order to aid interpretability.
Given that neither the “photography-based measures of feline skull conformation” or the online “owner-reported respiratory abnormalities” has been validated, results from the study can only be considered exploratory. Further, the study design has deliberately introduced biases by selecting the eldest cat and linking the questionnaire to specific websites and potential confounders are not controlled, making the study results unreliable and unsound scientifically. Moreover, the claims made on breed standards, cat welfare, and buyer’s awareness on brachycephaly are derived from inappropriate extrapolation of the study results and therefore are not considered valid. The authors should consider clarifying above points and publishing a revised conclusion.