Application of Clinical Metrology
Clinical Metrology tools may be used for veterinary studies evaluating an intervention or treatment; translational studies using naturally occurring disease in animals as a model of human disease; clinical veterinary patient management.
Prior to using a tool, one must know that it reliably measures what it is purported to measure and that it is able to pick up clinically significant changes in the population in which it is being used.
The process of validation is thus an important one, and consists of several stages:
- Item generation: The generation of the questions to be tested needs to be carefully undertaken, and have input from appropriate stakeholders.
- Using patients (painful and healthy populations)
- Using focus groups only
- Readability: it is important the instrument can be easily understood by those using it (i.e. the target population)
- Reliability: The stability of a tool examines the reproducibility of the tool administered on different occasions. When the tool is a questionnaire, its internal consistency is based on a single administration of the tool and represents the average of the correlations among the questions in the tool. Both internal consistency and stability must be proven before a questionnaire is to be deemed reliable.
- Face Validity: indicates whether, on the face of it, the tool appears to be assessing the desired qualities
- Content Validity: a judgment regarding whether the tool covers all of the relevant content
- Construct validity: Testing that is used when the tool is measuring something (a construct) that cannot be directly observed (pain, quality of life etc.). While the construct can not be directly seen, behaviors resulting from it can be observed. Obviously, it will be impossible to ‘prove’ that something that cannot be measured directly is being measured. Several approaches can be used:
- Hypothesized factors tested with factor analysis
- Discriminatory validity: Does the instrument discriminate between animals with and without the condition? Does it discriminate between different severities of the condition?
- Responsiveness of the tool to a treatment known to change what is being measured, or to a change in the condition over time
- Correlation to overall QoL
- Criterion or Concurrent Validity: the correlation of an instrument with some other measure – a measure accepted as the ‘gold standard’. Often, in the development of subjective assessment tools, the best approach is to use an accepted objective measure. Because pain cannot (yet) be directly measured, an objective surrogate measure can be used, e.g. an objective measure of activity if pain is expected to impact activity
Journal of Clinical chemistry and Laboratory Medicne