Background Machine learning (ML)–based clinical decision support systems (CDSS) are popular in clinical practice settings but are often criticized for being limited in usability, interpretability, and effectiveness. Evaluating the implementation of ML-based CDSS is critical to ensure CDSS is acceptable and useful to clinicians and helps them deliver high-quality health care. Malnutrition is a common and underdiagnosed condition among hospital patients, which can have serious adverse impacts. Early identification and treatment of malnutrition are important. Objective This study aims to evaluate the implementation of an ML tool, Malnutrition Universal Screening Tool (MUST)–Plus, that predicts hospital patients at high risk for malnutrition and identify best implementation practices applicable to this and other ML-based CDSS. Methods We conducted a qualitative postimplementation evaluation using in-depth interviews with registered dietitians (RDs) who use MUST-Plus output in their everyday work. After coding the data, we mapped emergent themes onto select domains of the nonadoption, abandonment, scale-up, spread, and sustainability (NASSS) framework. Results We interviewed 17 of the 24 RDs approached (71%), representing 37% of those who use MUST-Plus output. Several themes emerged: (1) enhancements to the tool were made to improve accuracy and usability; (2) MUST-Plus helped identify patients that would not otherwise be seen; perceived usefulness was highest in the original site; (3) perceived accuracy varied by respondent and site; (4) RDs valued autonomy in prioritizing patients; (5) depth of tool understanding varied by hospital and level; (6) MUST-Plus was integrated into workflows and electronic health records; and (7) RDs expressed a desire to eventually have 1 automated screener. Conclusions Our findings suggest that continuous involvement of stakeholders at new sites given staff turnover is vital to ensure buy-in. Qualitative research can help identify the potential bias of ML tools and should be widely used to ensure health equity. Ongoing collaboration among CDSS developers, data scientists, and clinical providers may help refine CDSS for optimal use and improve the acceptability of CDSS in the clinical context.
BACKGROUND Machine learning (ML)-based CDSS are popular in clinical practice settings but are often criticized for being limited in usability, interpretability and effectiveness. Evaluating implementation of ML-based CDSS is critical to ensure CDSS is acceptable and useful to clinicians and helps them deliver high quality health care. Malnutrition is a common and underdiagnosed condition among hospital patients that can have serious adverse impacts. Early identification and treatment of malnutrition is important. OBJECTIVE To evaluate the implementation of a ML tool, Malnutrition Universal Screening Tool (MUST)-Plus, that predicts hospital patients at high risk for malnutrition and identify implementation best practices applicable to this and other ML-based clinical decision support systems (CDSS). METHODS We conducted a qualitative post-implementation evaluation using in-depth interviews with registered dietitians (RDs) who use MUST-Plus output in their everyday work. After coding the data, we mapped emergent themes onto select domains of the nonadoption, abandonment, scale-up, spread, and sustainability (NASSS) framework. RESULTS We interviewed 17 of the 24 RDs approached (70.8%), representing 36.9% of those who use MUST-Plus output. Several themes emerged: (1) Enhancements to the tool were made to improve accuracy and usability; (2) MUST-Plus helped identify patients that would not otherwise be seen. Perceived usefulness was highest in the original site; (3) Perceived accuracy varied by respondent and site; (4) RDs valued autonomy in prioritizing patients; (5) Depth of tool understanding varied by hospital and level; (6) MUST-Plus was integrated into workflows and electronic health records; (7) RDs expressed a desire to eventually have one automated screener. Our findings suggest that continuous involvement of stakeholders at new sites given staff turnover is vital to ensure buy-in. Qualitative research can help identify potential bias of ML tools and should be widely used to ensure health equity. CONCLUSIONS Ongoing collaboration between CDSS developers, data scientists, and clinicians may help refine CDSS for optimal use and improve acceptability of CDSS in the clinical context. CLINICALTRIAL N/A
Introduction The emotional health of patients with IBD has been difficult to elucidate in routine IBD care, but is critical to medication adoption, adherence and self-management. Patients often are unsure how to communicate their preferences and concerns to their providers in ways that could directly inform shared decision making. Photovoice is an established research methodology used to give vulnerable patients a voice through alternative communication strategies, but has not been previously used in IBD. Aim Our goal was to determine the acceptability and feasibility of developing a communication tool using photovoice in an IBD clinic. The ultimate goal is to adapt Photovoice to facilitate doctor – patient communication around treatment and wellness goals in the clinic setting. Methods We recruited patients at a single tertiary care IBD center in 2019 to participate in a pilot Photovoice study. Patients received a digital camera, training on basic usage and 10 disease specific prompts focused on goals/strategies they used to manage IBD. For example, “What is the most important thing for your doctor to know about you?” Patients then participated in in-depth interviews where they shared the photos they took and described rationale for their photo choice. The interviews lasted approximately one hour and were audio recorded and professionally transcribed. Three analysts coded transcripts for themes using qualitative analysis software QSR NVivo 11. Subsequently, five physicians were recruited and also participated in in-depth interviews to gauge provider feasibility of incorporating Photovoice into clinical practice. Results Fifteen patients were enrolled, median age 28 IQR (24–40), 66% women, 86% white. Three patients (20%) identified as Hispanic and six (40%) identified as Ashkenazi Jewish. Fourteen transcripts were available for analysis (9 patient and 5 providers). A total of 87 photos were taken and reviewed with patients, with a subset of 15 photos reviewed with physicians. The general themes from patients were physical and psychological aspects of disease, logistical/practical aspects, and future with IBD. Physician response was overwhelmingly supportive of incorporating Photovoice into clinical practice and suggested several ways to incorporate: 1. As discrete parts of visits to foster goal-setting and identify patient priorities. 2. Displayed in the hallways of the clinic to foster community among patients. 3. As part of electronic medical records or as prompts in the waiting room to generate referrals to other resources like psychotherapy, social work and diet consults. Conclusions Photovoice is a feasible methodology for patients with IBD and acceptable to providers to use in a clinical setting. Photovoice may help providers identify patient concerns and tailor their communication and enhance approaches to shared decision making.
Objective To describe adaptations necessary for effective use of direct-to-consumer (DTC) cameras in an inpatient setting, from the perspective of health care workers. Methods Our qualitative study included semi-structured interviews and focus groups with clinicians, information technology (IT) personnel, and health system leaders affiliated with the Mount Sinai Health System. All participants either worked in a coronavirus disease 2019 (COVID-19) unit with DTC cameras or participated in the camera implementation. Three researchers coded the transcripts independently and met weekly to discuss and resolve discrepancies. Abiding by inductive thematic analysis, coders revised the codebook until they reached saturation. All transcripts were coded in Dedoose using the final codebook. Results Frontline clinical staff, IT personnel, and health system leaders (N = 39) participated in individual interviews and focus groups in November 2020–April 2021. Our analysis identified 5 areas for effective DTC camera use: technology, patient monitoring, workflows, interpersonal relationships, and infrastructure. Participants described adaptations created to optimize camera use and opportunities for improvement necessary for sustained use. Non-COVID-19 patients tended to decline participation. Discussion Deploying DTC cameras on inpatient units required adaptations in many routine processes. Addressing consent, 2-way communication issues, patient privacy, and messaging about video monitoring could help facilitate a nimble rollout. Implementation and dissemination of inpatient video monitoring using DTC cameras requires input from patients and frontline staff. Conclusions Given the resources and time it takes to implement a usable camera solution, other health systems might benefit from creating task forces to investigate their use before the next crisis.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.