C. Leveraging AI, Informatics, and Community-informed Methods to Advance Health Research: Data, design, and digital innovation in nursing science
Automate Creating, Customizing, and Optimizing Comorbidity Indices Using a Data-Driven AI/ML Approach presented by Chih-Lin Chi
Because individuals differ in severity, clinical studies typically use a comorbidity index to adjust for outcomes. On the other hand, because of its popularity, electronic health records are widely used to assess the quality of care. A subsequent question is how to adjust outcomes and control the severity. Although one may adjust outcomes using an existing comorbidity index, suboptimal adjustments may occur when applying it to different outcomes or patient subgroups. One can develop a new comorbidity index or modify an existing one to reduce this problem, but it takes time. This study proposes an Artificial Intelligence/Machine Learning model, called the Automatically Customized Comorbidity Index (ACCI), to generate a comorbidity index with EHR and a user’s outcome of interest. Specifically, ACCI customizes and optimizes the comorbidity index via prediction and optimization components. As an example, we use ACCI to create comorbidity indices for three outcomes of interest: statin-associated symptoms, statin therapy discontinuation, and statin days-supply. Here, we use random forests for prediction and a genetic algorithm for optimization. The results show that ACCI iteratively improved the comorbidity index’s predictive power and relevance to the outcome. In addition, the customized comorbidity indices show better performance in adjusting for these three outcomes than the baselines, the Charlson, and the Elixhauser comorbidity indices.
Elevating Voices in the Design of Complementary and Integrative Health Community Programs to Address Back and Neck Pain presented by Brent Leininger
Back and neck pain are common, often disabling, pain conditions and access to effective self-management interventions in the U.S. is often challenging. The purpose of this NIH R61/R33 project is to increase access to effective, complementary and integrative health self-management approaches through community programs. The goal of this presentation is to describe how both participant and community feedback was systematically sought and used to tailor programs and research processes to better address peoples’ needs. We used mixed-methods (quantitative and qualitative) data collection and analyses informed by an established behavioral model. Feedback was gathered in multiple ways at several points over the course of the pilot study (R61) leading to the full scale trial (R33). It included interviews to learn about pain-related lived experiences, surveys gathering opinions about the programs and research processes, and discussions seeking insights into the pilot results to inform further program optimization. Over 70 unique participants provided information. Interviews (n=28) identified barriers to past pain care included being negatively labeled and not feeling seen or heard. They also shared the desire for trustworthy information and quality pain care; choice and access to appropriate resources; and respectful partnerships with health professionals. Of the pilot study participants (n=51), >80% were satisfied with the programs, felt their pain improved, and liked what they learned. Importantly, >95% felt their beliefs were respected and staff worked in partnership with them. After viewing the results, participants provided important ideas for future implementation including involving health providers and employers, incorporating participant voices earlier in the process and others. Thinking more expansively about how participant and community partner views can be integrated iteratively over time can elevate their voices and improve scientific quality and relevance.
AI-assisted Online Intervention on Lung Cancer Screening among High-risk smokers: A pilot intervention study presented by Fang Lei
Background: Lung cancer remains the leading cause of cancer death in the United States, yet uptake of lung cancer screening with low dose computed tomography (LDCT) among eligible high-risk smokers remains low. Evidence suggests that limited knowledge, stigma, and health beliefs contribute to screening underutilization. Objective: This pilot study examined an online educational intervention aimed at improving high-risk smokers’ knowledge, attitudes, health beliefs, and behavioral intentions related to lung cancer screening. Methods: A pre–post intervention design was used. High-risk smokers, defined according to the U.S. Preventive Services Task Force (USPSTF) criteria, completed baseline questionnaires followed by five self-directed online educational modules delivered through REDCap. Post- intervention questionnaires assessed changes in lung cancer and screening knowledge, lung cancer stigma, health beliefs based on the Health Belief Model, and intentions regarding smoking cessation and LDCT screening. Lung cancer screening uptake was assessed via follow- up email three months after the intervention. Quantitative analysis included descriptive statistics and paired-samples t-tests; qualitative feedback was analyzed using constant comparison analysis. Results: Twenty-five participants completed the intervention (mean age = 60.9 years; 72% female; 88% current smokers). Significant improvements were observed across all major study outcomes. Knowledge scores increased markedly (3.76 to 8.60, p < .001). Lung cancer stigma decreased (25.52 to 19.16, p < .001). Health Belief Model constructs showed significant improvements, including perceived susceptibility, perceived benefits, cues to action, and self- efficacy, alongside reductions in perceived barriers and perceived severity (all p < .001). Intentions, importance, and confidence related to both quitting smoking and obtaining LDCT screening increased significantly Of the 22 participants who completed follow-up, 13 (59.1%) reported obtaining LDCT screening. Participant satisfaction with the intervention was high (mean = 18.32/20). Conclusions: Findings from this pilot study support the feasibility, acceptability, and preliminary effectiveness of an online educational intervention to promote lung cancer screening among high-risk smokers. The intervention improved knowledge, reduced stigma, positively influenced health beliefs, and increased screening uptake. Results provide a foundation for a larger-scale study and suggest that online educational platforms may be an effective strategy to reach geographically diverse high-risk populations and promote LDCT screening.