The use of Artificial Intelligence in the management of chronic pain: A Scope Review (2024)

Date created: | Last Updated:

: DOI | ARK

Category: Project

Description: The project aims to explore the literature on the topic of Artificial Intelligence and chronic pain, verifying through published studies what are the potentials, gaps and problems to be faced in this new and comprehensive field of study.PROCEDURES AND METHODSMethods Related to Article 1 (which will present the results achieved to address the specific objectives 1, 2, and 3 of this thesis): Artificial Intelligence in Chronic Pain Management: A Scoping ReviewWriting and Registration:The methodological steps and presentation of the scoping review results will follow the recommendations of the Preferred Reporting Items for Systematic Reviews and Meta-Analyses extension for Scoping Reviews (PRISMA-ScR) (MCGOWAN; STRAUS; MOHER; LANGLOIS et al., 2020; SARKIS-ONOFRE; CATALÁ-LÓPEZ; AROMATARIS; LOCKWOOD, 2021; TRICCO; LILLIE; ZARIN; O'BRIEN et al., 2018), which is an extension of the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) (MOHER; LIBERATI; TETZLAFF; ALTMAN et al., 2009) to be used for this type of review.This project is registered in the Open Science Framework (OSF) with the identification number: DOI 10.17605/OSF.IO/2GE63. Available at: https://osf.io/kv9hu/.Study Design:This is a Scoping Review, defined as: “a process of mapping key concepts in a research area, with its main sources and types of available evidence” (PETERS; GODFREY; MCINERNEY; MUNN et al., 2020); its main purposes are: to provide a deep analysis of the available literature, examining the extent, scope, and nature of research activity; determine the real need to conduct a systematic review on the subject; summarize and disseminate research results already conducted and identify gaps in the available scientific literature (ARKSEY; O'MALLEY, 2005). All methodological procedures followed the guidelines of the JBI Manual for Evidence Synthesis (AROMATARIS; MUNN, 2020), available at: https://synthesismanual.jbi.global.Review Question:What is available in the scientific literature regarding the use of Artificial Intelligence in chronic pain?Eligibility Criteria:Inclusion Criteria:Sample: individuals with chronic pain;Use of artificial intelligence techniques to optimize the management of individuals with chronic pain;Analysis of any type of outcome, preferably clinical outcomes;No restrictions based on race, gender, age, religion, etc.Exclusion Criteria:Research protocols or other publications that do not present original research results;Studies that have used any other method of data analysis or tool other than Artificial Intelligence.Bibliographic Search Strategy:A systematic search without time or language filters was conducted in the following databases: MEDLINE (US National Library of Medicine); LILACS (Latin American and Caribbean Literature in Health Sciences); SPORTDiscus; SCOPUS; Web of Science; CENTRAL; Science Direct; IEEE Xplore; Association for Computing Machinery Digital Library, and arXiv.The search phrase was performed by two independent authors, and disagreements were discussed and resolved with the entire group of authors. Boolean operators “OR” were used between synonyms and “AND” between descriptors. The descriptors, present in the Health Sciences Descriptors (DeCS) and the Medical Subject Headings (MeSH), “Chronic Pain” and its synonyms were used in combination with the terms “Artificial Intelligence,” “Machine Learning,” and their synonyms. The same descriptors were used for all databases, with the search phrase being specific for better clarity of the results and easy reproducibility. The references were managed using the Rayyan program, (OUZZANI; HAMMADY; FEDOROWICZ; ELMAGARMID, 2016) available at: https://www.rayyan.ai/.After performing the search in all databases, the studies were analyzed, and duplicates were excluded. After exclusion, two independent authors analyzed the studies by titles and abstracts, excluding those that did not fit the previously defined selection criteria. Disagreements between the authors were discussed and resolved with the entire group of authors.Types of Studies Included:No restriction on study design.Main Outcomes to Be Studied:Clinical outcomes: Pain, disability, quality of life, etc.Additional Outcomes:Adherence, usability, acceptance, etc.Data Extraction Strategy:Standardized tables were created for data extraction related to general characteristics, methodologies, results, and conclusions of each article. The data extraction was performed by two independent authors with experience in literature reviews. Disagreements in extraction were reviewed and decided by the entire group of authors.Variables Extracted from the Articles and Their Respective Categories:General Characteristics / Study MethodsAuthorYearCountryJournal’s Area of PublicationType of Chronic PainPain MeasurementType of CareType of StudySample SizeStudy ObjectiveTool Characteristics / DataTool TargetAI TypeAI Objective (classification/genetic exp./treatment/self-care/pain measurement)Data SourcesData TypesText/Image/AudioResultsConclusionAI Techniques UsedBayesian NetworkComputer VisionData MiningExpert SystemFuzzy ModelsNatural Language ProcessingRoboticsSupervised Machine LearningUnsupervised Machine LearningOthersStudy PurposeDiagnostic decision supportTreatment decision supportSpecialist referral supportPrognosis supportMedical records analysisKnowledge base constructionInformation extraction (structured or unstructured data)Provision of descriptive informationOther (open to use depending on the studies)Tool’s Target AudiencePatientHealth ProfessionalHealth ManagerCaregiverOthersTypes of Chronic PainMusculoskeletal chronic painChronic pelvic painFibromyalgiaMyofascial pain syndromeChronic cancer painChronic cervical painChronic low back painNon-specific chronic painOthersLevels of Care in the Unified Health System (SUS)Primary CareSecondary CareTertiary CareMethods Used for Literature Synthesis:To synthesize the information relevant to the study analysis, the variables were grouped and presented under the following themes: general characteristics/methods of the studies; tool characteristics/data; AI techniques used; study purpose; tool’s target audience; most prevalent types of chronic pain among the studies; intended level of care. Thus, aiming to facilitate the understanding of the theme proposed by the authors' narrative. Graphs were used to demonstrate the dimension of the studied variables, indicate patterns in the found results, and compare information relevant to the study context.