The main aim of the Collaborative for Research on Outcomes and -Metrics (CRΘM), established in 2008, is to change the way that biomedical data are analyzed. CRΘM brings methods and perspectives from psychometrics, neuropsychometrics and biometrics (those are the -Metrics) into clinical and biomedical research. This Collaborative capitalizes on experience and training in cognitive science and neuropsychology, measurement, statistics and biostatistics, bringing together expertise in computational statistics, clinical trial and study design and analysis, and psychometrics and other modeling skills of this type, belonging to affiliated scholars, to form a unique interdisciplinary research team.
The Collaborative is intended to bring together investigators interested in modeling -and applications of modeling (particularly in biomedical domains). The aim of CRΘM is to bring sophisticated modeling, in the support of measurement – and not limited to psychometrics- within and across biomedical applications. CRΘM is not a fee-for-service type resource, but is instead a full partnership-in-research type resource, where aspects of statistics and measurement in research (ongoing or new) can be explored.
Another way to achieve the main aim of CRΘM is to ensure that investigators know what they are actually implying when they use terms like “reliability” and “validity” (as well as when they don’t use such terms!). The “Θ” in our logo represents the common statistical representation of an estimator “theta”, particularly in the modern psychometric modeling sense that theta is the underlying (latent) trait of greatest interest. As a representative of the “Outcomes” in the Collaborative’s title, this underscores our commitment to expanding the idea of complex modeling beyond its origins in modern measurement and psychometric applications into clinical and biomedical research. Most CRΘM members are instructors or trainers in measurement and/or modeling, striving to strengthen abilities with, and critical thinking about, measurement arguments in biomedical and clinical contexts.
Measurement in clinical research: CRΘM drives innovation unlike any other data center in the US.
CRΘM is unique in the country for its emphasis on measurement – and not simply statistical modeling or psychometrics – in biomedical research. This focus on measurement in biomedical research is of particular concern in neurological and psychiatric outcomes, including but not limited to pain, function, cognition, mood, and neuropsychiatric symptoms and diagnoses. As they move towards clinical trials, investigators may seek proof of concept, pilot, or other clinical studies. Alternatively, investigators may have completed their pilot work and be ready to design a clinical trial for some intervention, but face operational and/or statistical challenges. These could include: identifying FDA-approved outcome variables, the best design of the study, facing small sample challenges and difficulties with covariates, difficult outcome quantification (e.g., in neurodegenerative disorders, slowing of progression, stabilization instead of decline), and consideration of co-primary endpoints.
Additionally, investigators might be considering new outcomes. The Collaborative for Research on Outcomes and -Metrics brings extensive psychometric and measurement analysis expertise to the identification, development, validation, and comparisons of outcome variables in clinical and biomedical research. Critical elements of the work by Dr.s Yang and Yumoto include expertise in the application and interpretation of modern test theory (sometimes referred to as psychometric or “modern psychometric”) models. These include Rasch, multi-facet Rasch, and more traditional 1-, 2-, and 3- parameter logistic models. These methods are widely used in educational contexts, and are also being utilized in the very large PROMIS effort undertaken by NIH (www.nihpromis.org) – these methods require very large samples of responses and assessment items with extremely specific characteristics (in terms of their targeting of the exact same skill, attitude or symptom so that they are exchangeable). With the background and experience of Dr.s Tractenberg and Yumoto, CRΘM can advise clinical trial and study investigators on the suitability and performance attributes of the PROMIS items (e.g., representing pain, fatigue, physical functioning, and caregiver burden) in the development of clinical trial and study design as well as CRFs using REDCap. The measurement expertise of CRΘM is longstanding, strongly integrated into the neurological research in our communities, and can propel innovations in clinical trials and studies to improve efficiency and interpretability.