Dr. Pavletic is senior staff clinician in the Office of the Clinical Director, Mr. Luckenbaugh is medical statistician in the Mood and Anxiety Program, Dr. Pao is deputy clinical director in the Intramural Research Program, and Dr. Pine is chief of developmental studies in the Mood and Anxiety Program, all at the National Institute of Mental Health in Bethesda, Maryland.
Disclosures: The authors report no affiliation with or financial interest in any organization that may pose a conflict of interest.
Disclaimer: The views expressed in this article do not necessarily represent the views of the National Institute of Mental Health, the National Institutes of Health, the United States Department of Health and Human Services, or the United States Government.
Please direct all correspondence to: Adriana J. Pavletic, MD, MS, 10 CRC, Room 6-5340, 10 Center Drive, MSC 1276, Bethesda, MD 20892-1276; Tel: 301-594-7386; Fax: 301-402-2588; E-mail: firstname.lastname@example.org.
• Assessments based on research volunteer-provided history are not sufficient in determining eligibility for protocols.
• Physical examination may discover psychiatric and/or medical disorder.
• Toxicology screen is often positive in research volunteers.
• Medical evaluation is equally important in healthy controls and anxiety patients.
Introduction: The importance of psychiatric screening of volunteers participating in research on mental illness is well established. Although psychiatric research frequently relies on subjects presumed to be free of medical conditions that affect nervous system function or safety of participants, little information exists on the value of medical screening in this population. This study describes findings on medical evaluations that potentially impact psychiatric research.
Methods: The authors conducted a retrospective analysis of medical evaluations in 476 consecutively referred healthy controls and 64 anxiety patients to determine the prevalence of conditions that resulted in exclusion from studies. All subjects had history and physical examination by a board-certified family physician and 37% of participants completed laboratory assessment.
Results: One-hundred ten (20%) volunteers were excluded. Exclusion rates were similar for controls and patients. The most common reasons for exclusion were psychiatric conditions (6.3%), positive toxicology screen (5.4%), abnormal liver function tests (4.5%), cardiovascular abnormalities (3.9%), positive viral markers including hepatitis C, hepatitis B, and human immunodeficiency virus (3.5%), anemia (2.5%), neurologic disorders (1.6%), and electrolyte abnormalities (1.0%).
Discussion: Medical screening identifies a relatively high rate of conditions in both healthy controls and anxiety patients that could impact on psychiatric research. A significant proportion of exclusions was found on physical exam, laboratory assessment, and toxicology screen.
Conclusion: These findings demonstrate the complementary nature of medical and psychiatric evaluations and underscore the need to develop further standards in medical screening procedures of volunteers in psychiatric research.
Previous reports demonstrate the importance of psychiatric evaluation1-5 and toxicology screening6-7 in individuals volunteering for mental health research. Research on mental illness typically attempts to recruit volunteers without medical conditions that might affect the functioning of the nervous system or safety of participants. However, in contrast to considerable work on mental health evaluation, few studies consider the value of comprehensive medical evaluation in this population.
Particular debate exists among mental health researchers regarding the need to perform physical exam and laboratory testing in volunteers participating in noninvasive studies such as functional magnetic resonance imaging (fMRI). Consequently, assessment of physical health often relies on a self report of medical history by potential volunteers. However, histories often fail to detect exclusionary conditions in volunteers participating in both psychiatric and medical research, possibly due to financial incentive.6-10 The aim of this article is to describe findings on medical evaluation that resulted in exclusions of volunteers from studies.
Five-hundred forty consecutive research volunteers, between 18–55 years of age (476 healthy controls and 64 anxiety patients) were medically evaluated from May 2003 through April 2005 to determine eligibility for one of nine protocols from four principal investigators. Volunteers were financially compensated for their participation. All protocols were approved by the National Institute of Mental Health (NIMH)-Intramural Research Program (IRP) Institutional Review Board. Two studies involved fMRI, four studies involved fear conditioning with electric nerve stimulation, and three studies involved fear conditioning and/or one-time medication administration.
These subjects were recruited when they contacted the NIMH-IRP. Recruitment methods for NIMH-IRP studies are modeled after those used throughout the various National Institutes of Health (NIH) IRPs, which, in turn, are modeled after those used throughout the medical community. Data on recruitment methods were not collected in this study. All subjects requesting participation were required to undergo an initial phone screen to determine potential eligibility. This initial screen typically led to exclusions among a relatively high proportion of potential subjects. Rates and reasons for these exclusions were not examined since the focus of the current report concerns rates of exclusion among subjects deemed to be eligible based on this initial screen.
Medical and Psychiatric Eligibility Criteria
All protocols required the absence of medical conditions and/or use of psychoactive medications that may affect the functioning of the nervous system or safety of participants. For healthy volunteers, inclusion criterion required the absence of a current Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition,11 Axis I mental disorder as determined either by the Structured Clinical Interview for DSM-IV Disorders (SCID), Non-Patient Edition,12 in seven studies or history and physical examination (H&P) in two studies. For anxiety disorder patients, inclusion criteria comprised current diagnosis of generalized anxiety disorder, social anxiety disorder, panic disorder, or specific phobia as determined by the SCID, Patient Edition,13 and study psychiatrist (D.S. Pine, MD).
Medical and Psychiatric Screening
Volunteers who passed initial standardized phone screens conducted by college level research assistants for healthy controls and mental health professionals for patients were subsequently evaluated in person. Clinical screening was completed by licensed mental health professionals (psychologists, mental health nurses, social workers) for the SCID and by a board-certified family physician (A.J. Pavletic, MD, MS) for the H&P and laboratory assessment. The order of in-person evaluation was determined by the availability of clinicians. Identification of exclusion criteria on an initial H&P or SCID precluded further evaluation. Thus, for example, volunteers initially receiving the SCID who met exclusion criteria were not medically evaluated and are not included in this report.
For healthy volunteers, screening procedures varied by protocol, including H&P in two studies, H&P and SCID in two studies, and comprehensive evaluations (H&P, laboratory assessment, electrocardiogram, and SCID) in five studies. For anxiety patients, all subjects received the comprehensive evaluation. Laboratory workup included complete blood count with differential, acute care panel (electrolytes, glucose, blood urea nitrogen, creatinine), hepatic panel (alkaline phosphatase, alanine aminotransferase, aspartate aminotransferase, bilirubin), thyroid-stimulating hormone, viral markers (HIV, hepatitis B, hepatitis C), qualitative urine drug screen (amphetamines, benzodiazepines, tetrahydrocannabiol, cocaine, opiates), and urine pregnancy test.
A family physician performed medical clearance of anxiety patients and determined eligibility of healthy controls based on all available information including H&P, SCID, laboratory assessment, and NIH medical record review of subjects who previously participated in NIH studies. The NIH maintains a comprehensive medical record for all potential research participants who have volunteered in any NIH-IRP study. Questionable cases were discussed with principal investigators (experimental psychologists) and study psychiatrist.
All studies used identical criteria to rule out conditions that might influence interpretation of study results. However, relative to noninvasive fMRI studies, medical eligibility criteria were more stringent in provocative fear conditioning studies and studies involving medication exposure, due to safety concerns. For example, liver function test abnormalities present on at least two occasions were exclusionary only in studies with medication exposure, and murmurs and mitral valve prolapse were exclusionary only in provocative studies with electric nerve stimulation. No subjects were excluded based on only one-time abnormal laboratory result, as one-time laboratory abnormality could be transient or caused by a laboratory error.
Fisher’s Exact test, χ2, chi-square, and t-tests, were used to compare healthy controls and patient volunteers on dichotomous, polychotomous, and continuous measures, respectively. Means and standard deviations are reported. Significance was evaluated at P<.05, two-tailed.
Anxiety patients were older (34±12 years) than healthy controls (27±8 years; P<.0001) and underwent more comprehensive evaluations. Laboratory assessment was completed in 37% of participants, ie, 89% of anxiety patients versus 30% of healthy controls (P<.0001). A SCID was administered in 66% of participants, ie, 97% of patients and 62% of healthy controls (P<.0001).
A total 110 of 540 subjects (20%) were excluded, 102 (19%) for medical or psychiatric reasons. Exclusion rates were similar for healthy controls and anxiety patients (Table).
The most common reasons for exclusion were psychiatric conditions (6.3%), positive toxicology screen (5.4%), abnormal liver function tests (4.5%), cardiovascular abnormalities (3.9%), positive viral markers including hepatitis C, hepatitis B, and HIV (3.5%), anemia (2.5%), neurologic disorders (1.6%), and electrolyte abnormalities (1.0%) (Table). Excluded subjects were older (mean=30.9, SD=10.3) than subjects accepted to protocols (mean=27.3, SD=8) (P<.001). As expected given differences in study criteria, exclusion rates were significantly higher in medication challenge studies (32%) compared to fMRI (18%) and fear conditioning studies (17%; X=13.57, df=2, P=.001).
Proportions of exclusions found during various methods of in-person evaluation are shown in the Figure. Forty-one subjects were excluded by history or history and SCID, while seven subjects were excluded by SCID only (Figure). Significant proportion of exclusions (59/107 or 55%) was detected by physical exam, laboratory testing, and NIH medical record review, ie, screening methods that rely on information beyond volunteer-provided history. Some examples of significant findings on physical exam include scarring from intentional self injury, very low body mass index (BMI) of 14.5, severe hypertension, tachycardia, conjunctivitis, and loud heart murmur probably indicating valvular heart disease.
The importance of laboratory testing for both healthy controls and patients is illustrated in the following examples. A 42 year-old healthy control had unremarkable H&P and SCID, but tested positive for cocaine and hepatitis C. Two anxiety patients tested positive for amphetamines. It is possible that their anxiety disorder was substance induced.
With the exception of one anxiety patient, all volunteers with positive toxicology screen denied any recent illicit drug use during phone screening, SCID, and H&P. For example, one healthy volunteer who had negative SCID had conjunctivitis on exam. He was drinking water from a large container during the interview. His toxicology screen was positive for tetrahydrocannabinol. Volunteers who tested positive for viral markers were significantly older (41±9 years) than those who were negative (21±9; P=.001).
NIH medical record review resulted in exclusion of seven volunteers whose H&P and SCID were unremarkable. For example, a healthy control who denied history of mental illness during the SCID and H&P had participated 1 year earlier in an NIMH treatment study as a patient with recurrent major depressive disorder. Medical record review of a 50-year-old healthy control who denied any medical problems revealed severe anemia with hemoglobin of 6.8 documented 6 months prior to current evaluation; she had applied for fear-conditioning study that did not require laboratory testing. An anxiety patient denied a history of substance abuse, but medical record review revealed a past history of polysubstance dependence. As this was not exclusionary in the study for which he applied, he underwent laboratory testing that later identified hepatitis C infection.
In eight healthy controls that were excluded for psychiatric reasons, the SCID revealed no Axis I diagnosis. However, observations during the H&P in concert with consultation with the study psychiatrist and principal investigators led to exclusion for psychiatric reasons. For example, exclusion followed the observation during physical exam of extensive scarring due to self injury. In another case, history identified attention deficit/hyperactivity disorder that had been previously diagnosed and treated by a psychiatrist outside the NIMH. None of these psychiatric conditions are routinely assessed by the SCID.
Ten subjects had more than one medical exclusion. For example, one healthy control had severe obesity with a BMI of 60, hypertension, and one-sided blindness. Another had history of meningitis with consequent hearing loss, hypothyroidism, and severe migraine headache treated with tryptan.
The current report is the first that specifically addressed findings on medical evaluation in healthy controls and anxiety patients who volunteer in research on mental illness. Although the study population in this cohort was young and relatively healthy, conditions were detected that could have a profound influence on the safety of participants and validity of research results, including severe hypertension, extreme weight disturbances, electrolyte abnormalities, viral infections, and positive toxicology screen. As in previous investigations,1-5 these results confirm that phone screens fail to identify sizable proportion of subjects who are ineligible for research on mental illness. For example, Shtasel and colleagues1 reported 47% of exclusions for medical and psychiatric illness but did not describe medical exclusions. Consistent with previous reports,1-7 the current cohort also displayed relatively high rates of psychiatric disorders and drug use in healthy controls. As the authors of this article did not include subjects who were excluded by the SCID prior to medical evaluation, the prevalence of psychiatric conditions in this cohort was significantly lower than in previous reports. Methods to improve the yield of eligible volunteers and increase the cost-effectiveness of the screening process have been previously reported2 and are not examined in this study.
Exclusion rates were different in various protocols due to differences in eligibility criteria and the extent of evaluation. For example, more stringent eligibility criteria and more extensive evaluation with laboratory assessment explain higher rejection rates in medication challenge studies.
Study results confirm previous observations that histories are often not reliable in assessment of eligibility of research volunteers, possibly due to financial incentive.6-10 Moreover, denial is common in some psychiatric conditions such as substance abuse and eating disorders.
Despite the fact that only 37% of subjects underwent laboratory testing and toxicology screen, 55% of exclusions in this study were found by physical exam, laboratory testing, or medical record review, ie, procedures relying on methods other than volunteer-provided history.
Study procedures in some research protocols required healthy controls to receive less extensive assessments than patients with anxiety disorders, and thus only 30% of healthy controls underwent laboratory assessment and toxicology screen. Gibbons and colleagues14 suggested the importance of screening healthy participants with a level of care equal to that applied to patients as inadequate screening of controls may adversely impact research results.
While healthy controls usually have no complaints, anxiety patients often present with a variety of physical symptoms. For example, dizziness, weakness, and palpitations may indicate anxiety, anemia, cardiac abnormality, substance abuse, or any combination of these conditions. Therefore, medical evaluation is equally important in patients.
Some findings on medical evaluation may represent complications of psychiatric disorders that had been minimized by volunteers during interview. For example, hypertension, tachycardia, abnormal liver function tests, infection with hepatitis B or C, or HIV may be consequences of substance abuse. One of the healthy controls whose blood pressure was 202/87 denied prior history of hypertension but admitted recent cocaine use after further questioning. Other findings in this cohort may represent manifestation of eating disorders, such as hypo-estrogenic amenorrhea, extremely low BMI, and electrolyte abnormalities. In cases where research volunteers may be motivated to conceal their problems, physical exam, laboratory assessment, and medical record review increase the sensitivity of in-person evaluation. However, some potentially serious preexisting medical conditions such as severe hypertension, valvular heart disease, and viral infections may remain unrecognized without a medical evaluation.
There are some limitations in this study inherent to its retrospective design. The variability in exclusion criteria and extent and order of in-person evaluation makes it somewhat difficult to interpret the results.
As some potential medical exclusions were not pre-specified, the research team reached decisions concerning eligibility on a case-by-case basis using all available information including subject’s age, other risk factors, and study procedures and invasiveness. However, it is impossible to pre-specify all potential exclusions. Moreover, there is insufficient knowledge and no consensus regarding many conditions and medications that may impact some forms of psychiatric research. Unlike psychiatric eligibility criteria, medical eligibility criteria and extent of medical evaluation are rarely discussed in psychiatric literature and deserve further study.
Medical screening identified a relatively high rate of conditions in both healthy controls and patients that potentially impacts mental health research. Perhaps most importantly, these findings demonstrate the complementary nature of medical and psychiatric evaluations and underscore the need to develop further standards in medical screening procedures for volunteers in psychiatric research. PP
1. Shtasel DL, Gur RE, Mozley PD, et al. Volunteers for biomedical research. Recruitment and screening of normal controls. Arch Gen Psychiatry. 1991;48(11):1022-1025.
2. Schechter D, Lebovitch R. Normal controls are expensive to find: methods to improve cost-effectiveness of the screening evaluation. Psych Res. 2005;136(1):69-78.
3. Huang DB, Koo H, Dougherty D, Hassan Y. Psychopathology among persons responding to participation as normal controls in behavioral research. Compr Psychiatry. 2003;44(2):83-87.
4. Halbreich U, Bakhai Y, Bacon KB, et al. The normalcy of self –proclaimed “normal volunteers.” Am J Psychiatry. 1989;146(8):1052-1055.
5. Bunce SC, Noblett KL, McCloskey MS, Coccaro EF. High prevalence of personality disorders among healthy volunteers for research: implications for control group bias. J Psych Res. 2005;39(4):421-430.
6. Struve FA, Straumanis JJ, Manno JE, Fitzgerald MJ, Patrick G, Leavitt J. Inadequacies of self-report data for exclusion criteria detection in marihuana research: an empirical case for multi-method direct examination screening. J Addict Dis. 2000;19(3):71-87.
7. Swerdlow NR, Geyer MA, Perry W, Cadenhead K, Braff DL. Drug screening in “normal” controls. Biol Psychiatry. 1995;38(2):123-124.
8. Apseloff G, Swayne JK, Gerber N. Medical histories may be unreliable in screening volunteers for clinical trials. Clin Pharmacol Ther. 1996;60(3):353-356.
9. Watson N, Wyld PJ. The importance of general practitioner information in selection of volunteers for clinical trials. Br J Clin Pharmacol. 1992;33(2):197-119.
10. Kolata GB. NIH shaken by death of research volunteer. Science. 1980;209(4455):475-476,478-479.
11. Diagnostic and Statistical Manual of Mental Disorders. 4th ed. Washington, DC: American Psychiatric Association; 1994.
12. First MB, Spitzer RL, Gibbon M, Williams JB. Structured Clinical Interview for DSM-IV-TR Axis I Disorders, Research Version, Non-Patient Edition. (SCID-I/NP). New York, NY: Biometrics Research, New York State Psychiatric Institute; 2002.
13. First MB, Spitzer RL, Gibbon M, Williams JBW. Structured Clinical Interview for DSM-IV-TR Axis I Disorders, Research Version, Patient Edition with Psychotic Screen (SCID-I/P W/PSY SCREEN). New York, NY: Biometrics Research, New York State Psychiatric Institute; 2002.
14. Gibbons RD, Davis JM, Hedeker DR. A comment on the selection of “healthy controls” for psychiatric experiments. Arch Gen Psychiatry. 1990;47(8):785-786.