Biomarker-Oriented Diagnosis and Therapy in Case of ADHD and Comorbidity

Final Report 2018
based on remarks by Dr. Andreas Müller, CEO Brain and Trauma Foundation Grisons


The project was supported by the Hirschmann Foundation between 2014 and 2018; it pursued the following goals:

  • Researching ADHD and its accompanying disorders over a lifespan of 7 to 55 years
  • Researching biomarkers under the consideration of neurophysiology, blood values and genetics in a naturalistic research design
  • Describing each study participant’s development over the course of two years
  • Researching the variables that primarily impact the development during those two years

The Brain and Trauma Foundation conducted the study "Biomarker-Oriented Diagnosis and Therapy in Case of ADHD and Comorbidity" in cooperation with the Psychiatric University Hospital Zurich.


Biomarkers

Evidence-based medical diagnostics makes use of a number of chemical and physiological parameters to connect biological evidence with the patient’s behavior and subjective perception. Conversely, today’s psychological diagnosis rarely integrates biological parameters. This gap could be closed by factoring biomarkers into the psychological diagnosis..

Biomarkers are biological variables that can reveal interdependencies between electrophysiological parameters and specific psychiatric diseases.

The current state of research indicates that there is not one single biomarker responsible for certain mental disorders. Biochemical analyses result in completely different results than neurophysiological methods.

DThe current state of research indicates that there is not one single biomarker responsible for certain mental disorders. Biochemical analyses result in completely different results than neurophysiological methods.

New approach on the Basis of Larger Samples

Employing an approved methodology, larger data samples of ADHA patients and healthy controls are collected according to scientific criteria, resulting in a relevant number of variables of biological information processing. The data are compared in a classification process until an optimal separation of the two groups is reached.

The aim is not to find individual variables, but combinations of variables that – as linked and complex functions – can separate the two groups as accurately as possible.

Values depend on the samples. If samples are not collected carefully, this will lead to mistakes and erroneous results. .

In order to obtain stable and reliable results, it is necessary to extensively test results on the basis of various classification methods.

Determined Sample Size

The determined sample size of approximately 500 people with ADHD as well as 250 healthy controls seems sufficient. Five follow-up check-ups of the ADHD group and three follow-up check-ups of the control group make it possible to re-evaluate the diagnoses several times.

Classification was done by a team of three experts who used 160 classification methods to test the data.

The achieved classifications each resulted in values of 80% sensitivity and specificity.

Application

Neuro-algorithms are pieces of jigsaw in the complex diagnostic process. Cognition, feelings, and behavior interact with neuro-biological conditions. Comprehending the biological data of a patient gives insights into the various brain functions.

On the basis of the collected data, classifiers were established that separate ADHD patients from healthy controls.

The most stable research method is regularized linear regression, as the data of new patients can be compared to the neuro-algorithms. Levels of correspondence reach from no correspondence (below 50%), slight correspondence (50-65%), medium correspondence (65-80%) up to high correspondence (over 80%).

Clinical Practice

Since spring 2017, the ADHD-Index has been clinically applied to many patients and with high caution. The initial results show a positive outcome in approximately 300 new patients. In over 90% of the cases, the results correspond with the patients’ own descriptions.

On the basis of 30 uniformly diagnosed patients with ADHD as their principle diagnosis, the ADHD-Index was first called into question. The diagnoses that had been made by a specialist were confirmed by the ADHD-Index at 100%:

  • there is a slight correspondence (probability between 50% and 65%) in 15 patients,
  • a medium correspondence (probability between 65 % and 80 %) in 11 patients, and
  • a high correspondence (probability between 80 % and 100 %) in 4 patients.)

Criticism

It is needless to say that the results are object to some criticism. The following counter-arguments have been raised:

  • humans are being replaced by machines
  • neuro-algorithms could be misused
  • neuro-algorithms have not been sufficiently tested yet

What's Next?

The ADHD-Index needs to be further tested and improved. A research group from Harvard, Boston, proposed to test the Index, which has been kindly accepted.

Subtypes need to be developed in order to reach better insights into possible treatment aspects.

Making deductions about positive as well as problematic courses of development must become possible.

Follow-Up Project

The follow-up project pursues the following goals:

  • Publishing the research results in adaption to each target group (doctors, psychologists, health insurances, insurances, patients, public).
  • Raising awareness and sensitivity in potential users by providing information and involving them interactively.
  • Promoting biomarker-oriented/biomarker-complementary diagnostics and evidence-based diagnostics in case of mental disorders, especially in the case of ADHD by involving users interactively.
  • Validating and adapting the concepts to the users‘ needs.
  • Promoting the evidence-based treatment of mental disorders.