I was quite excited when I heard about this paper in Biological Psychiatry a few days ago. It was hyped as a way to forgo the usual DSM approach and others and make an addiction assessment based more on the neuroscience of addiction. The basic dimensions for assessment are highlighted in the above diagram. The authors make a compelling argument in terms of what is needed in addition to the clinical criteria "that has provided a reliable foundation for the practice of addiction medicine." The clinical criteria that they are referring to are DSM criteria or basically problems and symptoms that are used to classify disorders from non-addictive use of the same substances. Even the most biologically based of these symptoms - craving, tolerance, and withdrawal vary widely across all individuals in the same diagnostic group. That variation is most likely due to biological complexity. The authors contend that there should be a way to examine that heterogeneity among the larger clinical divisions to get at pathophysiologically based subtypes. They suggest that the focus should be more on process than outcomes.
They use cancer as an example of the importance of specific etiology in the diagnosis and treatment of disease. In the case of the diagnosis and treatment of breast cancer, the BRCA1 gene is used to predict increased risk for breast cancer for a subpopulation of women with this diagnosis. Detection of HER2 protein overexpression can predict response to a monoclonal antibody (trastuzumab) that interferes with the HER2/neu receptor. All of this information is used within the existing clinical context and even then addition information about breast cancer would probably be useful. The review several similar initiatives in psychiatry and addiction and compares the ANA (Addictions Neuroclinical Assessment). The other examples include the Research Domain Criteria (RDoC), the Cognitive Neuroscience Treatment Research to Improve Cognition in Schizophrenia (CNTRICS), and the Impaired Response Inhibition and Salience Attribution (iRISA). Clinicians will note that none of these initiatives has gained a foothold in routine clinical practice. Only the CNTRICS has assessments across all 5 of the domains addressed by the ANA.
Looking at the major domain for assessment, the authors provide definitions and the rationale for their inclusion. They define executive function as "processes related to organizing behavior toward future goals". The reality is that executive function is really defined by convention, typically measures that are thought to reflect to reflect frontal lobe function like mental flexibility and set shifting. Addiction significantly limits the repertoire of these frontal mechanisms to maintain rather than critically assess the addiction. Impaired glutamatergic signalling to the stratum and extended amygdala and loss of top-down control are listed as the putative neuroscientific mechanisms. Incentive salience is described as "psychological processes that transform the perception of stimuli, imbuing them with salience and making them attractive." The underlying mechanism is given as "activation of mesocortical dopamine system". Familiarity with the reward and motivational system that is focused on the ventral tegmental area and its dopaminergic projections to the nucleus accumbens is at the heart of this system but it alson includes projections to the frontal cortex. During the initial phases of exposure to rewarding stimuli, the dopaminergic neurons will fire. As that process continues, anticipation of reward causes them to fire. That phasic dopaminergic activation leads to altered response to cue and noncue targets, craving, and heightened relapse risk. Negative emotionality is defined dysphoria and negative emotional responses to stimuli associated with addictive states. These states are often mistaken for an treated as depression. The ANA has instruments to assess hypohedonia. Brain stress and antistress systems are thought ot be involved with the latter contributing to negative emotionality.
When I look at the table of measures that comprise the ANA a couple of scenarios come to mind. The first is the omnipresent Attention Deficit-Hyperactivity Disorder (ADHD) diagnosis encountered in psychiatric practice. Most of these diagnoses are not made by psychiatrists. In the people who I reassess because they may have an addiction, when I ask them about the diagnosis, I am likely to hear: "My primary care doctor sent me to a psychologist and I had two hours of paper and pencil and computer testing." The problem is that there are no neuropsychological tests for ADHD, no matter how extensive. Most of the test battery would be for executive function - right out of the ANA and those tests are not necessary for the diagnosis. That led Barkley to come up with his own version of checklist symptoms that he thought matched the executive function deficits of the disorder better than the neuropsychological tests did. The second diagnosis is Alzheimer's dementia. Cortical dementias are based on higher cortical deficits, memory problems and the characteristic progression. An extensive test battery for the disorder is not indicated. I would argue that medical testing to rule out other causes is the single most important biomedical approach and that an extensive test battery would not add much.
In summary, there are several questions about the ANA. The first is whether it can ever be widely implemented in its current form. The total length of the test is 205 minutes on paper and three of the tests are based on neuroimaging. The authors estimate that it would take about 10 hours to complete and cost anywhere from $3000 to $5000 per individual. That alone restricts the ANA to urban areas where psychiatric clinics are well staffed and have access to neuroimaging and staff with the time and interest in complex diagnoses. That runs counter to a 30 year trend to ration detox and addiction services and largely move them off of medical campuses. It also runs counter to the collaborative care idea that suggests psychiatric staff can be marginally involved in primary care clinics that equate psychiatric diagnosis and treatment to a metric that can be completed in 5 minutes or less. Following that logic, the ANA might fly in these settings if it was 5 to 10 minutes long and would reliably lead to a prescription. A managed care organization (and they all are these days) will not be applying this kind of test to people with addictions. It is hard to determine how many people with addictions are seen and assessed by these organizations. The second question involves the cost-effectiveness argument applied to medicine. I am certainly averse to this argument, but all of the bean-counters seeing this will ask: "If we do all of these tests will it change anything? Current treatment of addiction is a crude proposition, but are there specific treatments based on the testing that will improve the process and outcomes of treatment?
From the pure egghead side of the equation, does the ANA go far enough in exploring the conscious state of the person with addiction? I find that very few assessments examine the cycle of euphoria and positive reinforcement and dysphoria and negative reinforcement best described in a statement by one of the coauthors of this paper:
"Addiction is a chronic relapsing syndrome that moves from an impulse control disorder involving positive reinforcement to a compulsive disorder involving negative reinforcement." - George Koob
As far as I know there are no more accurate statements about addiction that capture course, clinical phenomenology, and implicit neuroscience in one sentence. If that sentence was all that you knew about addiction, you could diagnose and treat most disorders by deriving your clinical interview from these first principles. The problem in the DSM world is that the interview ends up being a collection of impersonal markers that may of may not uniquely apply to the person being assessed. The ANA seems to take the encyclopedic approach to this problem. The intense euphoric state of early addiction is rarely explored. A lot of what constitutes a clinical evaluation is the tabulation of various adverse outcomes rather than the ongoing process.
Just adhering to the basic science of the situation are there easier and more straightforward approaches to executive function, negative emotionality, and incentive salience. I think that there might be. I am familiar with most of the measures suggested by the authors but the proper analysis requires a much closer look at all of the suggested metrics and how well they will discriminate between thousands of people with the same clinical diagnosis. Although I dislike using it in the same context, the PHQ-9 is a good example of a purported measure of depression that really does not discriminate and the authors include several similar measures in their list including the Beck Depression Inventory, the Beck Anxiety Inventory, the Fawcett-Clark Pleasure Scale, and the Toronto Alexithymia Scale. Degrees of freedom are important in thinking about the total number of items available to characterize the population versus large binary discrimination. A lot of these measures don't seem up to the task, but i am the first to admit that it would be useful to see research that focuses on that issue.
In the meantime, the take home message for any interested clinician is that the ANA is not ready for prime time - for a number of reasons - but stay tuned. There also seems to be an opportunity to come up with new assessments of the systems in question that are more efficient and a better complement to clinical practice.
George Dawson, MD, DFAPA
References:
