Why Knowing Healthy Controls Matters

Abstract

Although healthy volunteers often serve as controls or primary participants in neuroscience research, they are usually less rigorously screened than patients, which can have far-reaching implications for interpretation of study results. Although this issue has long been recognized, it is rarely discussed in the research literature. This article examines how the rigorous screening and characterization of healthy volunteers is key for quality research.

A 79-year-old previously healthy and extremely fit man developed progressive shortness of breath. His pulmonary function test (PFTs) results came back surprisingly normal. I suspected that his PFTs were falsely negative as he was compared to average (not extremely fit) 79-year-old men. Examining the report, I discovered that he was compared to 79-year-old women (Figure 1, left). Once the error was corrected his numbers changed, e.g., 93 (Figure 1, left) dropped to 76 (Figure 1, right). The interpretation changed from “normal” to “restrictive pulmonary disease” which was consistent with his clinical presentation.

Figure 1.

Erroneous Pulmonary Function Tests results from a wrong comparison (female gender, left) and corrected results (male gender, right)

In research, as in clinical practice, the interpretation of results depends on the characteristics of the comparison group. For example, whether patients with schizophrenia are found to have lateral ventricular enlargement depends on the choice of controls and not on patients (1).

In neuroscience research, healthy volunteers (HVs) are generally recruited to serve as controls for patients with psychiatric and neurological disorders or primary participants in studies of normal brain anatomy and physiology.

The characteristics of HVs are defined by the eligibility criteria of the research study. The purpose of the eligibility criteria is to assure the validity of research and safety of participants. Although there are no standardized eligibility criteria for HVs in neuroscience research, the most important requirement is a healthy nervous system. For practical reasons, the diagnosis of health is usually made by excluding pathology. Consequently, HVs generally cannot have psychiatric and substance use disorders, neurological disorders, sensory deficits (e.g., vision or hearing loss), brain injuries, brain infections, and medical conditions (such as hypertension) or medications (such as psychotropic medications) that affect brain anatomy or function.

In addition to eligibility criteria, the characteristics of HVs also depend on the methods of screening. However, compared to patients, HVs are frequently less rigorously screened which may affect research outcome (2, 3). For example, one multicenter comparison analysis found a greater variance in the cerebrospinal fluid biogenic amines among controls than among patients, possibly owing to the inadequate screening of HVs (2). The authors stressed the importance of screening healthy controls with a level of care equal to that applied to patients and recommended further study of the process by which controls are selected for research in psychiatry (2). Although proposed almost three decades ago, such studies remain rare in the research literature.

Consequently, the gap between the vetting of patients and HVs continues (3) In their 2009 analysis of 474 articles in Neuroimage, Mazziotta et al. reported that about 75% of human brain imaging studies had relied on a self-report or made no mention of the methods used for screening of HVs (3). Only 7% of studies verified their controls’ health status by performing a physical examination (3).

As can be seen from the following examples, closer evaluation frequently shows that many HVs who self-report as “healthy” should have been excluded (3–5).

Shtasel and colleagues reported that about half (155/312) of HVs who passed phone screening were subsequently excluded owing to present or past psychiatric, neurological, or medical conditions, which were discovered during in-person evaluations by physicians (4). Their study had stringent exclusionary criteria for HVs (e.g., current and lifetime history of Axis 1 disorder, first degree relative with Axis 1 disorder) and extensive screening process including psychiatric interview followed by a structured clinical interview, semistructured interview on first degree relatives, physical examination, laboratory testing, and a toxicology screen (4). We found exclusionary criteria in 20% (96/476) of young HVs (mean age 27⁺/− 8 years) who underwent history and physical examination and various combinations of additional screenings, including a structured clinical interview, toxicology screen, laboratory testing, and medical record review (5). Toxicology screen was positive in 6% (8/145) of our HVs (5). When developing an MRI atlas of a healthy human brain, Mazziotta et al excluded 48% (168/348) of self-reported HVs based on findings during the history and physical examination (3). Most of the excluded participants were over age 60 and had hypertension or neurological abnormalities (3).

As these examples demonstrate, volunteer- provided history is often not sufficient to determine the eligibility of HVs (3–5). Volunteer- provided history can be inaccurate for various reasons such as financial incentive or discomfort disclosing psychiatric disorder (5, 6). Additionally, some medical conditions that affect brain such as hypertension are often asymptomatic.

What are the appropriate screening methods for HVs participating in neuroscience research? There are no studies specifically addressing the appropriate screening methods in neuroscience research. Screening methods should balance study risks, scientific aims, characteristics of the study population (e.g., age), and resources. They should be carefully chosen as, in addition to costs, there are risks associated with screening such as false positives creating anxiety and additional testing (7). While self-reports may be appropriate for some research, studies involving quantitative objective measures should also use objective screening methods such as physical examination and a toxicology screen. Objective methods of screening have been shown to uncover not only medical but also psychiatric conditions (e.g., eating disorders, signs of self-injurious behavior, substance use) not revealed in the psychiatric history and a structured clinical interview (5, 6). Toxicology screening may have the added advantage of perhaps deterring HVs who use drugs from participating in mental health studies.

Although objective methods of screening increase the upfront cost, they are justified if they provide information that is important for research integrity and participants’ safety (6). Yet the fact remains that researchers, particularly non-clinicians, may not have the awareness, time, or resources for adequate screening of HVs. One solution may be the centralized screening, by a dedicated team of clinicians, to establish a well characterized list of HVs investigators can draw on for their research protocols. Providing such a resource has the potential to reduce the cost and improve the efficiency, safety, and quality of research.

The interpretation of clinical tests and research results depend on the healthy comparison group. Without adequate screening, the characteristics of HVs are unknown which may affect research validity unbeknownst to investigators. Therefore, knowing the characteristics of HVs is equally, if not more important.