Are Conventional Neuropsychological Assessments Subject to Bias?

Historically, researchers, pharmaceutical companies, and healthcare providers alike have relied on conventional neuropsychological assessments like the Mini-Mental State Exam (MMSE) and Montreal Cognitive Assessment (MoCA) to gain insights into how an individual’s neurocognitive abilities may be changing over time. However, these assessments are subject to several biases, resulting in noisy, highly variable results.

While conventional neuropsychological assessments can provide a decent indication of overall neurocognitive function, they fail to provide the degree of data granularity, ecological validity, and infrastructure required to assess small, intraindividual changes in function longitudinally. This becomes particularly problematic for pharmaceutical companies working towards providing clear, objective evidence of drug efficacy in clinical trials for neurological diseases such as Alzheimer’s disease and Parkinson’s disease. 

Think about it like this—if you’re trying to prove a diet is effective in helping individuals lose weight, but the diet will only help someone lose five pounds and the scale you’re using always varies by +/- five pounds, you likely won’t get the evidence you’re seeking, regardless of the diet’s efficacy. The same idea applies to assessing drug efficacy in clinical trials. Without granular, accurate, and reliable data, there is no way to draw objective conclusions about the impact of the drug on clinical trial subjects.

Let’s take a deeper look into why conventional neuropsychological assessments are subject to bias and introduce an affordable yet accurate and reliable approach to assessing neurocognitive function in clinical trials.

Biases in Conventional Neuropsychological Assessments

Broadly speaking, there are two categories of biases that conventional neuropsychological assessments are subject to: administration and the assessment itself, as detailed below.

Administration Biases

Administration bias in neuropsychological assessments largely stems from the fact that the assessments require a human to coach the subject and administer and score the assessment. This introduces human-to-human bias and consequently, human-to-human variability. Additionally, these assessments must be performed in an office, introducing the risk of environmental bias and potentially influencing the results.

Assessment Biases

One of the most widely recognized challenges with conventional assessments lies in the assessments themselves. They are highly subject to education, age, language, and cultural biases. The lack of consideration of these biases and the inability to adequately adjust results accordingly makes clinical trials with diverse populations quite challenging.

The activities performed in conventional neuropsychological assessments are not ecologically valid and only exercise a narrow cross section of neurocognitive domains. Primarily, the assessments evaluate aspects of cognition as opposed to a breadth of cognitive and functional domains. In other words, normal day-to-day human variabilities within the test takers themselves can significantly bias the assessment results.

The Need for Better Neuropsychological Assessment Tools for Clinical Trials

In addition to metrics from conventional neuropsychological assessments, clinical trials often include aspects such as protein levels via scans and other procedures to assess drug efficacy—however, protein levels in the brain may not be the best indicator of whether or not subjects are improving. This further drives the need for better neuropsychological assessment tools in clinical trials. To reduce the risk of administration and assessment biases, assessments need to:

• Implement standardized instructions that are ideally delivered digitally without human intervention.
• Utilize computerized and automated scoring and reporting methods.
• Be designed in such a way that the neurocognitive evaluation matches and aligns with neurocognitive states required to complete Activities of Daily Living (ADLs).
• Exercise a breadth of both cognitive and functional domains.

With that said, assessing ADLs is arguably the most accurate and robust indicator of how a patient’s neurocognitive function is changing with drug or therapeutic intervention.