Page 49 - Clinical Small Animal Internal Medicine

17

VetBooks.ir

3

Using Data for Clinical Decision Making

Philip H. Kass, DVM, MPVM, MS, PhD

Department of Population Health and Reproduction, School of Veterinary Medicine, University of California, Davis, CA, USA

The advent of the electronic medical record was accom- a typically brief period of time to obtain an adequate
panied by a new‐found feasibility for veterinary medical total sample size, which are called “cross‐sectional stud-
practitioners, either at individual hospitals or through ies.” The second involves those with data on groups of
multiinstitutional collaborations, to gather, analyze, patients with some defining inclusion criteria collected
interpret, and publish their clinical data. While it is still over a specified period of time, allowing measurement of
common to read individual case or case series reports in a temporal change in patient characteristics and health
medical journals, electronic data and searchable data- status, which are called “longitudinal” studies. The latter
bases extend the ability of veterinarians to design more can be further subdivided into those that are interven-
sophisticated studies that go beyond the bounds of being tional (the clinical analog of experimental), where the
purely descriptive. Such design enhancements allow the investigators have control over the treatment(s) of pri-
measurement of differences and associations between mary interest the patients receive, and observational
two or more factors of interest, and in some cases lead to (nonexperimental), where the investigators can passively
estimation of measures of effect that under assumptions or actively record, but not manipulate, the treatment(s)
of unbiasedness can lead to causal inferences. of primary interest. The focus here on treatments arises
Using clinically derived data to make informed medi- from the nature of clinical research: any treatment in a
cal decisions is a central tenet underlying the practice of clinical setting should be pursuant to the best interests of
evidence‐based medicine, an evolving form of scientific the patient and owner, and should not be any lesser qual-
scholarship that seeks to bring objectivity and elucida- ity of treatment than the accepted standard of care in
tion to the implementation of medical investigation, and veterinary medical practice.
that exists at the confluence of medicine, basic sciences, As noted, analyses of treatments are not restricted to
applied mathematics/biostatistics, epidemiology, and interventional longitudinal studies, although it will be
causal inference. Using data analytically is only one shown that observational longitudinal studies of treat-
component of evidence‐based medicine, and cannot be ment effects are prone to potentially severe biases.
dissociated from equally important others. While col- Longitudinal studies also need not be restricted to study-
laboration is all but essential in ensuring that these ing treatment effects: intrinsic characteristics of patients,
respective component specialties are well represented, and how those may be associated with clinical sequelae,
of most proximate importance to veterinarians seeking may instead be a primary objective of a researcher.
to advance beyond the hypothesis phase of their clinical
research is a fundamental understanding of hospital‐
based epidemiologic study design, for this is the study Hospital‐Based Cross‐sectional Studies
type that determines the data collected, and in turn how
it should be analyzed and interpreted [1]. Cross‐sectional studies are conducted when a clinician
The most common study designs employed in clinical wants to establish an inventory of disease burden,
veterinary research will be considered in subsequent patient characteristics, and exposure to factors in a pop-
sections of this chapter. These will be broadly partitioned ulation. A shelter veterinarian, for example, may be
into two temporal types. The first involves data measure- interested in knowing the crude proportion of a shelter
ments taken on an assemblage of patients all present at a population at any defined time point that has clinical
point in time, or on a succession of patients enrolled over signs of upper respiratory disease. This proportion is

Clinical Small Animal Internal Medicine Volume I, First Edition. Edited by David S. Bruyette.
© 2020 John Wiley & Sons, Inc. Published 2020 by John Wiley & Sons, Inc.
Companion website: www.wiley.com/go/bruyette/clinical