text
stringlengths 1
1.75k
|
|---|
Illness |
How a particular person experiences disease, condition, or symptoms, including the collected effects on the person’s existence and identity such as biological, psychological, social, and economic outcomes. (See: Disability, Health-Related Quality of Life) |
Imagination |
In part, imagination is simply “the specific ability to produce and to decode images,” as media theorist Vilém Flusser put it. This can be further reduced to the realization of mental images, in whatever modality (thinking of a word, sound, place, gesture, plan, system, etc). This act can be solitary or collective. Imagination encompasses, as sociologist Ruha Benjamin lists: dreams, dreaming, ideas, ideology, stories, speculation, playing, poetry, myths, visions, narratives. Benjamin notes the strong role of imagination in the field of sociology, as “the capacity to link individuals’ personal problems with broader social processes.” Imagination is a line of flight, a fulcrum, a tool for both repair and new creation. Writer Imani Perry situates imagination prior to practice: “Imagination doesn’t erase nightmares, but it can repurpose them with an elaborate sense-making or troublemaking.” (See: Analysis, Apperception, Imaginary, Interpretability) |
Imaginary |
Imaginary concerns social articulations of what is possible: the many ways we express what we can think, what we think might happen, and what we think can be done. Imaginary is not in opposition with notions of truth, it is part of how truth thrives. Imaginaries are collective projections of a desirable and feasible future, per Ruha Benjamin’s succinct definition. For poet and writer Édouard Glissant, each culture has its own particular imaginary which is expressed in “aIl the ways a culture has of perceiving and conceiving of the world.” Psychoanalyst Jacques Lacan placed the imaginary with clinical specificity—as all that is available to the senses—alongside the symbolic (what language can do) and the real (the unrepresentable aspects of reality). To be imagined is to be, as Glissant puts it, “conceivable in transport of thought” . Sociotechnical Imaginaries, a concept developed by Sheila Jasanoff and Sang-Hyun Kim among others, describes how visions of scientific and technological progress carry with them implicit ideas about public purposes, collective futures, and the common good. (See: Common Sense, Discourse, Ideology, Imagination, Latent Space) |
Impairment |
“A special kind of limit.” While Impairments have in the past been used negatively in both legal and theoretical frameworks, as the medical basis for social constructs of disability, impairments are not in themselves good or bad. Having an impairment creates a re-routing of intention through “a productive distortion of an ability.” where “The ability itself might be real or imagined.” (See: Crip, Disability, Distortion, Imaginary, Noise) |
Incommensurable |
Fundamentally different. A characteristic of two or more entities that cannot be compared because they do not share a common standard, point of reference, or measurement. Incommensurability is a significant concept in mathematics, philosophy, and ethics: where translation is not possible, analysis is built on interpretation. |
Inference |
The action of predicting what is likely, based on evidence, ideas, methods, or analogies. Inference is something both people and machines do. It is a natural way of thinking. It has been formalized as the basis of both statistics and AI. Different kinds of inference include: inductive (starting with evidence), deductive (starting with ideas), statistical (starting with evidence and a set of methods), abductive (jumping between ideas and evidence by guesswork), or analogical (finding similarities between different situations). (See: Artificial Intelligence (AI), Model) |
Information |
“All instances where people interact with their environment in any such way that leaves some impression on them [...] These impressions can include the emotional changes [and] can also reflect complex interactions where information combines with preexisting knowledge to make new understandings.” To inform is to give form to something. Information is raw data that has been worked on: abstracted, annotated, analyzed, measured, compared, tracked, interpreted, synthesized, modeled, represented, etc. (See: Data, Endpoint, Model, Noise, Signal) |
Instrument |
An instrument is an extension of a person’s senses and actions that reliably represents their intentions, even as it adapts to new settings. |
Interpellation |
Interrupted, hailed, called, compelled to respond or self-identify in a particular way or shape to a particular prompt. To make yourself into a pattern. (See: Discourse, Disidentification, Ideology, Imaginary) |
Interpolation |
In mathematical terms, placing a new point between two (or more, depending on dimensionality) discrete points, smoothing a curve. In musical terms, not a sample, not a cover, but a quotation, like carrying a familiar tune). (See: Homotopy, Latent Space) |
Interpretability |
In the context of artificial intelligence, a model can be called interpretable if humans interacting with the model can experience firsthand how input features influence the output. An interpretable model’s inner mechanisms are understandable to humans. All interpretable models are also explainable, in that the justification for outcomes can be communicated; however, not all explainable models are interpretable, e.g. deep learning models. (See: Artificial Intelligence, Deep Learning, Explainability, Hermeneutics, Model) |
Intersubjectivity |
A shared understanding; a concept that emphasizes the way perceptions, experiences, and interpretations are shaped by interactions with others. Intersubjectivity is invoked in situations where personal experience, even sense of self, is co-created by interactions within social contexts and dynamics. (See: Common Sense, Discourse) |
Intervention |
Any action taken to relieve the burden of disease, or to improve a patient’s health status. Interventions may refer to medical procedures, preventive measures, therapeutic treatment, behavioral modification, and/or diagnostic tests. (See: Minimal Clinically Important Difference (MCID), Patient-Reported Outcomes (PROs), Patient-Reported Outcome Measures (PROMs), Treatment (Tx)) |
Large language model (LLM) |
A generative AI model, typically with billions or trillions of parameters, that has been trained on excessively large amounts of text or multimodal data in order to perform natural language tasks, including understanding and generating human-like language. (See: Generative AI, Latent Space, Retrieval Augmented Generation (RAG)) |
Latent Space |
A compressed representation capturing the underlying structure and patterns of some original data as embeddings. The spatial relation between embeddings in latent space can provide insight into how different samples are related in terms of their features, both explicit and implicit. In generative AI, a model learns to encode data into latent space, then generate new data by sampling from the latent space. New data sampled from the latent space will be unique, but will maintain a coherent representation of the patterns or features of the original data. Manipulating or coordinating between different latent spaces is key to the functionality of large language models, multimodal modals, and translation models, among others. (See: Compression, Data, Embedding, Generative AI, Homotopy, Interpolation, Manifold, Variational Autoencoder (VAE)) |
Life-Writing |
Life-writing covers “everything from the complete life to the day-in-the-life, from the fictional to the factional. It embraces the lives of objects and institutions as well as the lives of individuals, families and groups. Life-writing includes biography, autobiography, memoirs, letters, diaries, journals, anthropological data, oral testimony, eye-witness accounts, biopics, plays and musical performances, obituaries, scandal sheets, and gossip columns, blogs, and social media.” (See: Exposome, Hypomnemata, Narrative Medicine, Patient-Generated Health Data (PGHD)) |
Linguistic validation |
Adjusting for how changes in language produce new patterns in data. |
Longitudinal |
Describes data sampled at periodic intervals over a timespan (See: Data, Momentary, Retrospective, Sampling) |
Machine Learning |
A subset of artificial intelligence that describes a range of algorithmic approaches to learning from data. Learning, in the context of machine learning, means: given a set of data, a task to be performed on the data, and a measure of how well the task has been performed, perform the task well on both the original (i.e. ‘training’) data and new (previously unseen) data. In practice, this means creating a model of patterns and relationships in the original data, identifying a ground truth for how the model should perform the task, and adjusting the model’s parameters to minimize errors on the original data while maintaining the capacity to generalize to new data. (See: Artificial Intelligence (AI), Data, Model, Performance) |
Manifold |
A kind of model that generalizes mathematical ideas about space. A manifold can be any number of dimensions: for instance, a curve is a one-dimensional manifold, and a surface is a two-dimensional manifold. Latent space, as the projection of learned feature representations in many AI models, is a many-dimensional manifold of as many dimensions as the representation holds. (See: Homotopy, Latent Space, Model, Topos) |
Minimal Clinically Important Difference (MCID) |
The threshold at which a change in one’s health becomes meaningful. Find by asking the patient: would you consider repeating this intervention if you had the choice to make again? (See: Intervention, Patient-Reported Outcomes (PROs), Patient-Reported Outcome Measures (PROMs)) |
Model |
An abstract representation that tries to capture what is meaningful about a thing. (See: Compression, Homotopy, Latent space, Manifold, Sampling) |
Momentary |
In the context of research, a momentary context refers to a brief window of time in which interventions happen, tests or measurements are conducted, or reports are solicited: i.e. in the moment. (See: Longitudinal, Retrospective, Sampling) |
Multimodal |
In the context of AI, learning relationships between different representational modes such as image, sound, text, video, and so on. (See: Artificial Intelligence (AI), Model) |
Narrative Medicine |
Understanding health in the context of life. (See: Anamnesis, Emplotment, Exposome, History (Hx), Hypomnemata, Life-Writing, Patient-Generated Health Data (PGHD)) |
Natural Language Processing (NLP) |
How computers come to understand what humans mean, as when people use language to express themselves to one another. (See: Artificial Intelligence) |
Neurodivergent |
Of an individual’s particular experience with non-normative affective or cognitive processes. |
Neurodiversity |
A representation of the given (“actual” in Deleuzean terms) spectrum of non-normative affective or cognitive processes. |
Noise |
Used in distinction to signal (as a desirable source of information)—as excess, impairment, background, mask. Qualitatively, noise is a way of locating meaningful difference that is affective, non-representational and non-discursive. (See: Discourse, Distortion, Impairment, Incommensurable, Information, Non-Representational, Signal) |
Normal |
The force of coherence to dominant patterns. For philosopher and historian of science Georges Canguilhem, normal, in the context of health and illness, is the way an organism adapts and remains the same, maintaining stability in response to its changing environment. In this usage, normal is not fixed or static, it is a dynamic process of adjusting and regulating, where internal and external factors combine to shape what’s normal. (See: Common Sense, Discourse, Ideology, Interpellation, Model, Non-Representational) |
Non-Representational |
In non-representational research, the emphasis is not solely on representing or mirroring an external reality, but rather on engaging with the complexities and multiplicities of human-environment interactions, embodied experiences, and the relational nature of knowledge production, as well as lived experiences, interactions, and affective dimensions that may not be easily captured through traditional forms of representation. Informed by post-structural and post-humanist thought, non-representational research challenges stable identities, hierarchies, and boundaries, emphasizing the fluidity, contingency, and relationality of social and material phenomena. Non-representational research methods may be performative and/or collaborative, focusing on the role embodied, sensory, and relational experiences play in shaping knowledge and understanding. (See: Disidentification, Model, Normal) |
Outcome |
In research as in clinical care, an outcome is a variable being measured, such as a particular survey score, or lab result. It is data captured by means of an instrument (anything from a writing prompt to a medical device); it can be structured or unstructured, qualitative or quantitative. (See: Data, Endpoint, Instrument) |
Overhealing |
Scar tissue, etc |
Pain |
“An unpleasant sensory and emotional experience associated with, or resembling that associated with, actual or potential tissue damage.” |
Parataxis |
Producing poetic insight by placing things side by side |
Pathography |
Writing about illness (See: Anamnesis, Emplotment, Life-Writing) |
Patient Advocacy |
Representing the interests and rights of patients on individual, community, or systemic scales. Advocacy includes making sure patients’ needs are met, from access to care to education, decision-making, and ethics. (See: Patient-centered care (PCC)) |
Patient-Centered Care (PCC) |
A practical ethics of empowering patients to become active participants in their own care, emphasizing the importance of understanding patients’ perspectives and values, involving them in decision-making, and offering tailored approaches that suit individual patient’s needs. (See: Minimal Clinically Important Difference (MCID), Patient Advocacy, Translational Medicine) |
Patient-Generated Health Data (PGHD) |
Health-related data (e.g., health history, device data, and patient-reported outcomes (PROs) that are created, recorded, or gathered by or from patients, family members, or other caregivers to help address a health concern or promote health. The use of PGHD offers a unique opportunity to fill in gaps in information and provide a more comprehensive picture of ongoing patient health for use during care, resulting in potential cost savings and improvements in health care quality and outcomes, care coordination, and patient safety. |
Patient-Reported Experience Measures (PREMs) |
A kind of Patient-Reported Outcome Measure (PROM), PREMs are focused on the quality of a patient’s experience in terms of the health services they receive, their interaction with the health care providers or automated health systems and tools. PREMs often take the form of a satisfaction questionnaire. (See: Patient-Centered Care (PCC), Patient-Reported Outcome Measures (PROMs)) |
Patient-Reported Outcomes (PROs) |
Patient-Reported Outcomes are timely records of a patient’s experience of illness—their inner thoughts, bodily sensations, and social experience. PROs are raw data coming directly from the patient, without interpretation of the patient’s response by a clinician or anyone else. PROs are one kind of Patient-Generated Health Data (PGHD). (See: Apperception, Endpoint, Intervention, Minimal Clinically Important Difference (MCID), Outcome, Patient-Generated Health Data (PGHD), Patient-Reported Outcome Measures (PROMs), Self-report) |
Patient-Reported Outcome Measures (PROMs) |
A PROM is a standardized tool: a survey, instrument, scale, or single-item measure used to assess particular PROs, such as symptoms, behaviors, or functional abilities, as perceived by the individual, obtained by directly asking the individual to self-report. (See: Endpoint, Intervention, Minimal Clinically Important Difference (MCID), Outcome, Patient-Reported Outcomes (PROs)) |
Phenomenology |
The study of what makes experience and action possible. (See: Apperception, Study) |
Plain language |
Plain language is a way of writing that is more accessible. It uses smaller words and shorter sentences. This helps people (with intellectual and developmental difficulties, English language learners, etc.) to understand the main ideas more clearly. (See: Accessibility, Accommodation, Easy Read) |
Poetics |
The makingness of things and events; the “revelation and distillation of experience” that gives names and shapes to ideas and imaginaries, so they can be thought. (See: Hermeneutics, Imaginary, Interpretability) |
Precision |
How close multiple predictions are to each other—consistency and exactness. (See: Accuracy) |
Predictive Decision Support Intervention (DSI) |
“Technology intended to support decision-making based on algorithms or models that derive relationships from training or example data and then are used to produce an output or outputs related to, but not limited to, prediction, classification, recommendation, evaluation, or analysis” (89 FR 1192)(See: Analysis, Automated Decision Making (ADM), Model) |
Prompt Engineering |
Prompt Engineering is the set of opinions on how to use language to effectively interact with a model of language. (See: Large Language Models (LLMs), Retrieval Augmented Generation (RAG)) |
Quality of Life (QoL) |
(See: Health-Related Quality of Life (HRQL)) |
Realist Synthesis |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.