The best established and validated, computerized, is the Achenbach Test, google under Thomas Achenbach, this organization sells both the forms and the evaluation program. Much tested for the last 35 years

I would also consider screeners such as the BRIEF, parts of the Nepsy II or the SDQ  to assess ADHD.

The BRIEF is one of the most reliable standardized assessment tools for Parent /Teacher observers, since it takes into account this diagnostic profile as an outcome of disrupted cognitive executive functioning. I typically rely upon this tool in tandem with results from the Connors-3 and the CAB (Clinical Assessment of Behavior). I suggest keeping in mind utilization of a thorough, initial developmental history alongside these measures, as well as a request for narrative observer feedback from independent raters (e.g. teachers, etc.).    

Hi all. Thank you for your responses. I am aware of the BRIEF and other questionnaire/interview measures, and I always use the Conners and SDQ, but I was looking for something that does not depend on parent or teacher report, something like the ADOS that will allow an examiner to press for ADHD behaviours and will yield an examiner-rated ADHD severity score. It seems like the TOF is the closest to this. Thanks for your input!

I'm not sure if this is exactly what you had in mid, but the children's version of the Test of Everydy Attention is pretty good: http://www.pearsonclinical.com/psychology/products/100000480/test-of-everyday-attention-for-children-the-tea-ch.html

Im a clin psych in private practice and the one we use is the Conners 3 rating scale that can be done with the parent teacher or self report.  It is scored online for 3 dollars from MHS assessments  

I do both ADHD and ASD assessment and administer the ADOS-2 when doing ASD assessments.  The short answer for whether there is an observational system which "presses" for attention problems is "no". 

By the diagnostic criteria of the DSM for ADHD, using an observational system such as the ADOS which "presses" for attention problems would be against the idea of diagnosing ADHD. Attention problems are much more common than ASD behaviours (which are relatively rare/uncommon) and you would not know whether behaviours exhibited by a child are due to a primary attention problem or other explanations (for instance, anxiety).  When you are conducting an assessment, your own observations are a good indicator of what the child is like and in a way, the assessment situation is a "press" for attention problems occurring (e.g., boredom due to monotony, reactions to challenging situations, persistence/lack of persistence etc).  In the end run, it is the ratings of parents and teachers which determine the diagnosis on rating scales because ADHD is (according to DSM) a behavior disorder.   Basing your diagnosis solely on neuropsychological testing (such as the Continuous Performance Task) does not effectively discriminate those with ADHD or other conditions (refer to Barkley's ADHD handbook for a good review on this).

Wondering if anyone else has an opinion on this.

BASC-3 SOS is something to check on. It is a reliable observation instrument with multiple uses.

Speaking as both an educator and a neurobehavioral geneticist focusing on the evolutionary rise of developmental disorder...

You would have to first establish how it is to be distinguished from classification as "gifted" (which does not require demonstration of high I.Q.). Traits considered typical for ADHD have little to no variance from those used to identify the creatively gifted, especially when the child is twice exceptional (e.g. gifted w/ dyscalculia or dyslexia), or developing in a high risk environment.

What child psychologists/psychiatrists trained to look for deficits call "ADHD" (i.e. a pathology worthy of treatment) is quite often what better educators discover is (yes) a form of psychopathology classified as "gifted" under special education - and then address accordingly. In truth, there is very good data from fMRI imaging that better determines what someone might wish to call "ADHD" than the various screeners / inventories / interviews can hope to provide.

The continuous performance tests can be used and recently we have used the computerised QB test which comes in a version for 5-12 year and over 13 year olds.

the ADIS-C contains a short list of ADHD symptoms to be checked in an interview with child and parent separately

I'm not sure this answered has been fully answered, although posted in 2015. I would like to highlight some important issues in your question, Elisabeth:

1. The focus on the informant. The scales mentioned in previous answers in this thread are not clinician-rated (as you requested), but are all parent- or/and teacher-rated.

2. The focus on rating behaviour based on direct observations, instead of using a period of time leading up to the rating as the basis of the rating (which is often the last week, the preceding six weeks or the preceding 6 months).

3. The focus on a test situation including a structured observation specifically developed to assess ADHD-symptoms, with tasks or/and materials that will "challenge" the child's ability to overcome problems due to inattentiveness, hyperactivity and impulsivity (equivalent to ADOS, which is a structured observation test, designed to exhibit symptoms of ASD). 

The BASC-3 SOS (student observation scale) was suggested. This one comes close, as it assesses behaviour as directly observed, but it is actually a teacher-rated scale of behaviour in the classroom. Still, it could be a relevant addition to the regularly used parent- or teacher rated questionnaires.

The ASEBA system was also suggested and this system does include a some scales, which can be relevant to this question:

First, the TOF (Test Observation Form; ages 2-18 years), in which an examiner or an observer can rate the behavior of a child or an adolescent during IQ-testing. So, this fulfils the requirement of the informant being a clinician (not parent or teacher), but the IQ-testing may not be the best situation to assess the child's behaviour - although it is a situation which children and adolescents with ADHD often find challenging - it was not specifically designed for this type of observation. 

Second, the SCICA (Semistructured Clinical Interview for Children and Adolescents; ages 6-18 years) is a interviewer-based assessment of the child's reported difficulties in different areas of daily life, using open-ended questions and does include a direct observation of the child (SCICA - Observation Form), in which an assistant (or the interviewer) rates 121 items assessing observed symptoms of anxiety, depressive disorder, ADHD, ODD/CD and language and motor problems. Again, the informant is correct, but the test situation is not. This approach is similar to that of other semistructured interviews with children or adolescents, such Kiddie-SADS, in which the interviewer also rates observed symptoms and compare these observations to the answers provided by the interviewee. However, perhaps the SCICA - Observation Form could be applied to a structured situation (not an interview), similar to that of ADOS.

Third, the DOF (Direct Observation Form, ages 6-11) seems to fit almost all the requirements. This is a standardized form for rating observations of children’s behavior in classrooms, at recess, and in other group settings. During 10-minute periods, observers write narrative descriptions of children’s behavior and rate on-task behavior at 1-minute intervals. It is rated by a clinician, and based on a direct observation in different (mainly school-based) settings, very focussed on rating on-task behaviour in daily-life settings. However, it is not intended for observation of adolescents (only age 6-11 years) and it does not use a designed structured situation for the observation. The TOF seems to have good discriminative validity and practical utility (see www.ncbi.nlm.nih.gov/pubmed/20802814 , www.ncbi.nlm.nih.gov/pubmed/20802813 and www.ncbi.nlm.nih.gov/pubmed/20706918). 

I want to mention two instruments which can used for direct observation of ADHD-symptoms in the assessment of pre-school children, the Child Solo Play (www.ncbi.nlm.nih.gov/pubmed/11314565) and the Global Impressions of Parent-Child Interactions (GIPCI-R), which is a direct, semistructured observation schedule developed to evaluate parent-child interactions (APA, Handbook of Psychiatric Measures, 2008, pp. 309–342). Both of these two instruments are used in monitoring effects of non-pharmacological interventions in this age-group (www.ncbi.nlm.nih.gov/pubmed/27076496). And although not instruments developed for school-aged children (which is what you asked for) they otherwise fulfill the requirements and could perhaps be used as shaping molds to develop instruments for older children. I also want to mention the Manchester Inventory for Playground Observation (MIPO) (www.ncbi.nlm.nih.gov/pubmed/21831591). This instrument is not focused on assessing ADHD symptoms, but social functioning in school-aged children, and may also be useful in monitoring effects of ADHD-interventions.

To my knowledge there is no assessment tool for ADHD in school-aged children, equivalent to the ADOS used in the assessment of ASD. However, we may not always need direct observations to make the correct ADHD-diagnosis. In fact, a recent paper examining the existing literature suggests that, when compared to a gold standard diagnosis, rating scale information from parents and teachers, are equally accurate. So, obtaining teacher information may not always be necessary, if good parental information is available (www.ncbi.nlm.nih.gov/pubmed/28271921).

Furthermore, some children assessed for ADHD, in which reactive attachment disorder (RAD) is an important differential diagnosis, likely represent the cases in which a direct observation is highly beneficial, aiding clinicians to better distinguish between RAD and ADHD (www.ncbi.nlm.nih.gov/pubmed/21257287).

https://www.ncbi.nlm.nih.gov/pubmed/11314565

https://www.ncbi.nlm.nih.gov/pubmed/20802814

https://www.ncbi.nlm.nih.gov/pubmed/20802813

https://www.ncbi.nlm.nih.gov/pubmed/20706918

https://www.ncbi.nlm.nih.gov/pubmed/27076496

https://www.ncbi.nlm.nih.gov/pubmed/21831591

https://www.ncbi.nlm.nih.gov/pubmed/28271921

https://www.ncbi.nlm.nih.gov/pubmed/21257287

Thank you, Soren! This is most helpful. At the moment we are using actigraph measures of activity levels alongside the Conners and parent interviews and aim to test how actigraph measures relate to ADHD symptoms in children of a wide ability range. These at least should help with having an objective measure of hyperactivity. I will check out those references, especially the parent-child interaction. Many thanks!

Joseph, good point that observing symptoms of inattention, hyperactivity and impulsivity may not be equally straight forward. However, I don't agree complete that hyp/imp symptoms are always easier to identify for an observer, although sometimes they are.

Still, from my clinical experience subtle signs of at least some types of inattention can easily be observed, simply when talking to children in the office (and it would be similar in a structured situation - like the ADOS). During conversation, I might notice very brief and small changes in the child's ability to keep on track, during conversation. Or they don't really hear my question (mind being elsewhere) or are easily distracted by something they see outside the window. Or I notice trends in the child's attention level, during the consult (changing from being more on track in the beginning and less so at the end) and often also signs of exhaustion (becoming more tired at the end of our conversation, yawning, changing posture, sliding down on the couch). If they make a drawing, while I talk to the parents, sometimes they will do it too fast (make careless mistakes). And also how well they manage to stick to the overall plan for the consultation, as I lay it out for them in the beginning of our talk or have difficulties finding back to my office after they've been to the restroom (planning and organisational skills). They may also leave behind clothes, shoes, toys, phones etc., when leaving my office after the consult (losses things).

Similarly, some types of hyperactivity may not be clear to an observer. Restlessness is mainly an internal feeling. Also the criteria "on the go" may simply be due to feeling uncomfortable about having to be still for extended time. Both of which are not always easily observable. And many children with ADHD develop coping strategies to avoid acting impulsive, like for instance putting their hand in front of their mouth, to avoid talking too much, or stop themselves when just about to interrupt, which is observable, but not as clear impulsive behaviour as mentioned in the diagnostic criteria.

I mainly work with children and adolescents with ADHD, so I'm biased and probably tend to be especially focussed on evaluating the symptoms of that specific disorder :)