Guidelines for Conducting CFS Research Studies

The CFIDS Association of America encourages the research community to adopt consistent standards for the study of CFS research in an effort to improve comparability of findings and integrity of the field. This was one of the top recommendations arising from the April 2011 NIH-sponsored ME/CFS State of the Knowledge Workshop.

The following document represents an effort to describe factors that may improve consistency in study methodology. Applicants to the Association's Grants Program are strongly encouraged to follow these guidelines to the best of their ability. Deviations from these guidelines will not be excluded from consideration, but should be well-justified within the proposal.

Hypotheses
Research applications must have clear, identifiable, testable hypotheses, as well as clear indications for how each hypothesis will be addressed by the study.

If the study includes the evaluation of a particular treatment, the applicant should provide a clear rationale for testing the treatment, including pilot study data, clinical experience and background research. Applicants should describe the methods for providing consistent treatment delivery and identify, and perhaps quantify, possible confounds of the therapy. Attention should be paid to the degree to which control or comparison groups may be contaminated by awareness of the therapy or the likelihood that concurrent self-medicating can affect outcomes.

Defining Cases
The international standard for defining CFS in a research setting is the 1994 consensus case definition published in the Annals of Internal Medicine. Although clinicians may use less rigorous standards to diagnose CFS in practice, in research studies the 1994 definition should be the minimum means by which CFS cases are classified. Particular care should be taken to exclude other causes of symptoms. Subjects should be described as completely as possible, with data collected on duration of illness, type of onset, severity, functional status and comorbid conditions.

A clinical consensus definition for ME/CFS was developed in 2003 by Carruthers, et al. Applicants for Association funding are strongly encouraged to evaluate subjects using this definition as well as the 1994 research definition. The ME/CFS clinical definition (sometimes referred to as the Canadian concensus definition) can be accessed here: http://www.iacfsme.org/Portals/0/pdf/CanadianCaseDefinition.2003.pdf. Data analysis should reflect any differences that may be identified based on classification of subjects using this definition.

Selection of Controls
Controls should be representative of the population under study. They should be well-matched by age, gender, activity level, etc. So-called "contact controls" - persons who have regular contact with CFS patients - are usually not the ideal control group and should not be selected unless there is a specific rationale for doing so, which should be thoroughly described in the research plan.

Stratification and Subgrouping
Characterization and partitioning of study subjects into well-defined subgroups, per the 1994 CFS definition (or 2003 clinical definition), is necessary. Criteria for subgrouping must be well-defined and appropriate to the research questions. Examples of subgrouping criteria include age, duration of illness, comorbid conditions (fibromyalgia, migraine, irritable bowel syndrome, sleep disorders, depression, anxiety, etc.), type of onset, cognitive impairment, biologic markers, severity of illness, current and pre-illness body-mass index and functional status. Other factors may relate to the research focus, such as presence, degree or absence of hypocortisolism in neuroendocrine studies or orthostatic intolerance in autonomic nervous system studies. In addition, these factors should also be considered in data analysis.

Definition of Terms
Study terms should be clearly defined, especially those terms used differently across disciplines or with multiple meanings that could be misunderstood without clarification. For example, rather than stating that a study will evaluate "stress," it would be clearer to name the specific stressors (exercise, infection, etc.) that will be tested.

Nomenclature
The practice of interchanging the terms "chronic fatigue" and "chronic fatigue syndrome" is confusing and should be abandoned. Please use the following definitions when writing about CFS and related conditions:

• Chronic fatigue: the symptom of extreme tiredness or exhaustion which interferes with major life activities for six or more months
• I diopathic chronic fatigue: chronic fatigue which does not meet criteria for CFS, either because the fatigue severity or symptomatic criteria has not been met.
• Chronic fatigue syndrome: an illness characterized by chronic fatigue (as defined above) accompanied by four or more of the following symptoms: impaired memory or concentration, sore throat, tender cervical or axillary lymph nodes, muscle pain, multi-joint pain, new headaches, unrefreshing sleep, and post-exertion malaise. The symptoms must not be explainable by a medical or psychiatric illness.
• ME/CFS: This term is used somewhat interchangeably with CFS, although it also refers to the condition defined by the 2003 clinical definition by Carruthers et al. The U.S. Secretary of Health and Human Services is considering a recommendation to formally use ME/CFS in place of CFS, although the impact on definitions was not considered as part of the recommendation made by the DHHS CFS Advisory Committee.

Testing and Measures
Biological and psychological testing should be directly relevant to the aims of the study. The CFIDS Association of America will not fund testing unless it is clearly related to the hypotheses and assesses the outcome measures outlined in the study design. Instruments may need to be modified to correct for the symptom overlap of CFS with other conditions, including affective disorders. Applicants are urged to clearly explain the purpose, relevance and hypothesized outcomes of each test they propose to use in a study. Data should be provided on the validity and reliability of tests and measures and the populations on which they were normed.

Multidisciplinary Approaches and Collaborations
CFS is a systemic illness characterized by dysregulation in a number of highly integrated modulatory systems. CFS provides an ideal opportunity to foster exciting research collaborations in what might appear to be unrelated disciplines. For example, the integration of research findings in areas as divergent as cardiovascular physiology, neuroendocrinology, immunology, infectious disease, psychiatry, sleep physiology, epidemiology and genetics. Heretofore, CFS research has been largely limited to research within single disciplines. The Association encourages the formation of multi-disciplinary teams to develop a fertile, broad-based and far-reaching approach to elucidate the pathophysiology and treatment of this disorder. The Association has also prioritized studies that leverage (and/or strengthen) existing resources and infrastructures, including but not limited to the SolveCFS BioBank for subject enrollment and sample collection.

Study Strengths and Limitations
Applicants should clearly describe the strengths and limitations of the proposed study, in terms of study population, assessment tools, statistical or other analysis, etc. Applicants should demonstrate that they have considered potential limitations or weaknesses and attempted to address them within the study to the best of their ability, within the parameters of available funding and existing knowledge.

Power
Many CFS research studies published to date employ small numbers of subjects. Within the limits of budget and other factors, researchers must ensure that studies are adequately powered to address primary and secondary hypotheses and reduce the risk of errors related to attempts to draw conclusions based on too few subjects.

Statistical Analysis
Statistical tools should be used correctly and efficiently. Tools should be specified in the proposal. The most effective statistical tools for the nature of the study, including non-parametric statistics, stratification and regression modeling, should be employed, as appropriate.

Investigators must be sensitive to the issue of whether their results are practically meaningful, in addition to statistically significant. Small, but not stastically meaningful differences may be useful clues for future research, even if not for clinical application, and should be presented in appropriate context. Investigators are encouraged to analyze data post hoc in an exploratory manner to determine whether clinical subgroups have different outcomes.

For more information contact:

Suzanne D. Vernon, PhD
Scientific Director
The CFIDS Association of America, Inc.
PO Box 220398
Charlotte, NC 28222-0398
Tel.: 704-364-0016
Fax: 704-365-9755
E-mail: sdvernon@cfids.org

Updated July 11, 2011