Are existing psychiatric and cognitive assessment scales adequate for use in clinical trials of PD treatments? Do better PD-specific scales need to be developed, and how can they be validated?

No. And perhaps in some instances: yes. The development of PD specific assessment instruments should not be an aim in itself, but serve clear clinical and/or research purposes. In the most common clinical settings there are at present no convincing arguments for the development or use of PD specific assessment instruments. My arguments to support this stand are in part based on common sense and in part on scientific evidence.

1. The average general practitioner, neurologist, neuropsychiatrist or neuropsychologist, will see patients with a wide variety of neurological diagnoses, such as PD, stroke, MS, epilepsy, traumatic brain injury, the various types of dementia, etc. The same arguments that are used in favour of development of PD specific rating scales can be used to argue for the development of scales that are specific to any of these and other neuropsychiatric disorders. This would leave the clinician with a large number of scales, all addressing the same symptom domain, and all meant for use in a small population with a specific disease. In this way it will be difficult to gain experience with the ins and outs of all these different scales and their interpretation. One single scale for use in a large range of patients would be preferable, if the clinimetric properties of this scale across all these diseases are acceptable. This would facilitate the use of the instrument in clinical practice and probably have more impact on the quality of care. An exception to this general rule may be specialized movement clinics in large or academic hospitals or research settings, where a large number of PD patients are seen. In these settings PD specific assessment scales could potentially be of use IF it can be shown that they clearly have better clinimetric properties then non-disease specific scales.
2. However, up until now, no PD-specific scale has proven to perform better clinimetrically than non-disease specific assessment instruments. This holds true for psychosis, depression, anxiety, apathy and cognitive deterioration (Fernandez et al, 2008; Schrag et al, 2007; Leentjens et al, 2008; Leentjens et al, 2008).
3. Moreover, for many neuropsychiatric symptoms, the clinimetric properties of available assessment scales have not been assessed to a sufficient degree, or not been assessed at all, as is the case in anxiety and apathy rating scales (Leentjens et al, 2008; Leentjens et al, 2008). In the absence of knowledge about how existing scales perform in PD patients, it is not logical to start developing new scales.
4. Finally, the arguments often used to defend the development of PD specific scales are not valid. The overlap of non-motor symptoms of PD with neuropsychiatric symptoms is often mentioned as a limitation of existing assessment scales for psychopathology, such as in depression, although there is no evidence to support this. Scales with a lot of such overlapping items do not perform worse then scales without those items, and in the case of depression, some overlapping (somatic) symptoms even prove to be more specific to the psychiatric condition then to PD (Schraget al, 2007; Leentjens et al, 2003).

The same arguments used against the development of PD specific scales in clinical practice can be used to defend this view in research settings. For most neuropsychiatric syndromes, existing scales either have adequate enough clinimetric properties or have not been studied enough to draw conclusions. There may be one exception to this general statement. The commonly used cognitive screening instruments, such as the MMSE or CAMCOG, are often criticized because they are biased towards cortical neurocognitive symptoms, whereas cognitive decline in PD tends to be more characterized by subcortical symptoms. Although for individual subcortical cognitive symptoms a large range of specific neuropsychological tests are available, detection and assessment of these symptoms in clinical practice will probably improve by the use of a more balanced brief cognitive screening instrument that captures symptoms of ‘subcortical' cognitive decline and dementia. But here again: this is preferably not a PD-specific instrument, but an instrument that would capture such symptoms irrespective of the diagnosis across a range of clinical conditions.

Within the past several years a number of reports have highlighted the high frequency of cognitive deficits in Parkinson's disease. This has been accompanied by the realization that instruments for measuring cognitive function tailored for Parkinson's disease for use in both the clinical and research settings are needed. Importantly, most existing cognitive assessment instruments were developed for use in Alzheimer's disease, but the cognitive profile is distinct from that seen in Parkinson's disease. Several tools have been created that could be suitable for a research setting or for dedicated neuropsychological clinical assessments of PD patients, including the SCOPA-Cog (Marinus, 2003), the CAMCOG-R (Athey, 2005) and a brief neuropsychological protocol reported by Ong and colleagues (Ong, 2005). However these are too long for use in regular clinical follow-up of PD patients. I would argue that it is in assessing and/or optimizing scales for use in the clinical setting that the greatest work is still needed.

For clinical use, shorter instruments are needed that are sensitive to the cognitive deficits seen in PD. Interest in the Montreal Cognitive Assessment as a useful screening tool has led to at least three comparisons of this instrument to the commonly used MMSE with promising results (Gill, 2008, Nazem, 2009, Zadikoff, 2007) and it has already been incorporated into assessments of PD subjects involved in longitudinal cohort studies (PostCEPT). The MoCA was developed for detecting mild cognitive impairment in the general population and is available in many languages (www.mocatest.org). Other tools that may be useful in the clinical setting as screening instruments include the Mini Mental Parkinson (Mahieux, 1995) (developed in French and not, to my knowledge, translated into English) and the Parkinson Neuropsychometric Dementia Assessment (PANDA, reported using a German group of subjects) (Kalbe, 2005). Addenbrooke's Cognitive Examination has also been explored for use in PD as a brief (15-20 minute) assessment. (Reyes, 2009) Each of these short instruments has theoretical advantages in screening cognition in Parkinson's disease over the MMSE, given that they assess executive function which is clearly relevant in Parkinson's disease.

Although some performance characteristics have been reported, further work is needed to define the optimal cutoff scores and associated sensitivity and specificity for a diagnosis of both dementia and mild cognitive impairment using these brief instruments, and their responsiveness to change over time. It would be ideal to have direct comparisons of these instruments to decide which is/are best for both diagnosis and follow-up of dementia and mild cognitive impairment in patients with Parkinson's disease.

From the previously posted responses by Dr. Leentjens and Dr. Marras, it seems that what is needed isn't a collection of scales developed specifically for PD, but scales that have been validated in PD patients. Having shorter, more manageable scales is also key to helping ensure broader use by physicians who aren't necessarily movement disorder specialists. 

Both Drs. Leentjens and Marras are currently working to do just that in the clinic. Are there others?

At today's "Management of neuropsychiatric non-motor aspects" session at MDS, chaired by Paolo Barone and Marcelo Merello, discussions touched on the scales to use when measuring cognitive abilities in clinical practice.  Participants referred to a review by Kulisevsky and Pagonabarraga, 2009 as a good general resource.  In the discussion,  David John Burn asked participants to first think about why the scale is being used in order to help determine which is most appropriate.  A discussion ensued on the necessity of screening every patient for cognitive impairment using a formal scale. Is this something that can be evaluated through regular patient-clinician interactions or is it necessary to conduct formal assessments?

An interesting discussion arose at a recent MJFF Awardee Meeting regarding assessing mild cognitive impairment in Parkinson's disease (PD-MCI). Cognition decribes a very complex set of functions, but how do you choose which functions and endpoints to assess when assessing PD-MCI? Do you include executive function, memory function, certain components of both? How do you account for highly intelligent study subjects, in which measures of decline are still in the 'normal' range?