Development of an individualized pain prediction tool in pediatrics: Identification of risk factors, patient reported experience and outcome measures, and data capture tools for pre- and post-operative use (Preprint)

Abstract

BACKGROUND The perioperative period is a data-rich environment with potential for innovation through digital health tools and predictive analytics (e.g., to optimize patients’ health with targeted pre-habilitation). While some risk factors for postoperative pain following pediatric surgery are already known, the systematic use of pre-operative information to guide personalized interventions are not yet widespread in clinical practice. OBJECTIVE Our long-term goal is to reduce the incidence of persistent post-surgical pain and long-term opioid use in children by developing personalized pain risk prediction models that can guide clinicians and families to identify targeted pre-habilitation strategies. To develop such a system, our first step was to use patient-oriented research methodology, leveraging parental/caregiver and clinician expertise, to identify risk factors, outcomes, and relevant experience measures, as well as data collection tools for a future data collection and risk modeling study. METHODS We conducted virtual focus groups with participants recruited at a tertiary pediatric hospital; each session lasted approximately one hour and was composed of clinicians and/or family members (people with lived surgical experience and parents of children who had recently undergone a procedure requiring anesthesia). Data were analyzed thematically to identify potential risk factors for pain, as well as relevant patient-reported experience and outcome measures (PREMs and PROMs) that can be used to evaluate the progress of postoperative recovery at home. This guidance was combined with a targeted literature review to select tools to collect risk factors and outcomes information for implementation in a future study. RESULTS Twenty-two participants (12 clinicians and 10 family members) attended 10 focus group sessions; participants included 12/22 (55%) persons identifying as female, and 12/22 (55%) were under 50 years of age. Thematic analysis identified five key domains: 1) demographic risk factors (e.g., age, sex, weight), 2) clinical characteristics (e.g., length of hospital stay, procedure type, medications, pre-existing conditions), 3) psychosocial factors (e.g., anxiety, depression, medical phobias), 4) PREMs (e.g., patient and family satisfaction with care), and 5) PROMs (e.g., nausea and vomiting, functional recovery, return to normal activities of daily living). Participants further suggested desirable functional requirements (e.g., use of standardized and validated tools, longitudinal data collection) and delivery modes (e.g., primarily electronic, parent proxy and self-report) for tools that can be used to capture these metrics, both in hospital and following discharge. Established PREMs/PROMs questionnaires, pain catastrophizing scales, and substance use questionnaires for adolescents were subsequently selected for our proposed data collection platform. CONCLUSIONS This study established five key data domains for identifying pain risk factors and evaluating postoperative recovery at home, as well as the functional requirements and delivery modes of selected tools with which to capture these metrics both in hospital and after discharge. These tools have been implemented to generate data for the development of personalized pain risk prediction models.