A number of approaches have been taken in developing rating scales. Good examples are found in scales designed to rate depression; some are developed from ‘gold standard’ diagnostic criteria. For example, the OARS Depressive scale² is derived from depressive symptoms listed by DSM III diagnosis. Other scales have been developed from already existing rating scales. A good example of this is the Carroll Rating Scale (CRS)³ which was developed from the Hamilton Depression Rating Scale and the Beck Depression Inventory. Being aware of the developmental source of an individual instrument will inform choice in the context of the purpose and population in which the instrument is to be applied. For example, there are significant and important questions relating to the validity of DSM criteria in the diagnosis of depression in people with other disorders such as Parkinson’s disease¹¹. This generates the obvious question as to whether rating scales derived from these criteria are suitable for, or have been validated in these specific populations. Likewise, those rating scales that have been developed from other rating scales pose the same questions; have the source rating scales been validated in the target populations? Has there been subsequent validation of the specific rating scale in the relevant population? An obvious example is the use of rating scales for rating depression in patients with dementia. Another lies in the problem of rating depression in people with significant physical illness, where some of the physical illness symptoms, such as appetite loss; may be the same as some of those that are rated as depressive symptoms by the scale.

The same issue relates to rating scales designed to rate cognitive symptoms. The Mini Mental State Examination⁴ is probably one of the most commonly used cognitive assessment instruments. Its original purpose was to differentiate organic from functional disorders and was designed to quantify change in cognitive state. Despite its popularity, the instrument does have its critics as a screening instrument⁵. There are two important issues worth highlighting in relation to its use. Firstly, the instrument does not and should not be used to allocate a diagnosis. When used in the context of an acute medical ward, it is as likely to pick up cognitive deficit in people with acute confusional states as a consequence of physical illness as it is to identify cognitive problems associated with dementias. Secondly, the instrument fails to assess all cognitive domains, in particular, failing to address specific deficits relating to frontal cortex. This problem is reflected in the adaptation of the instrument over many years of use.

A good example of this is the development of the ACE-R¹², which was derived from the Mini Mental State Examination but includes a section addressing frontal lobe assessment.

Many of the above issues fall within the over-all heading of ‘validity’. Proving the validity of an instrument is seminal to its development. The concept addresses a number of issues in which the characteristics of the instrument in terms of its performance are addressed within defined populations. Hence, an instrument may be valid in one setting (or in rating symptoms in one type of population) but may not be valid in another. In testing the validity of an instrument; the degree to which it adequately probes the specific issues that are being rated is tested. Its ability to differentiate between people that are known to be dissimilar is tested, and finally, the instrument is examined in terms of the demonstrating that certain explanatory constructs account to some degree for the performance of the test. From a practical point of view it is critical that the choice of an instrument is informed by evidence of validation in populations similar to those that are intended to be interviewed.

Many rating scales have been developed as potential ‘case’ finding instruments. Usually a rating scale will generate a numerical ‘score’, often purported to represent severity of a particular set of signs or symptoms. Common examples include the Geriatric Depression Scale.