Identifying High-Need Primary Care Patients Using Nursing Knowledge and Machine Learning Methods

Abstract

Background: Patient cohorts generated by machine learning can be enhanced with clinical knowledge to increase translational value and provide a practical approach to patient segmentation based on a mix of medical, behavioral, and social factors. Objectives: To generate a pragmatic example of how machine learning could be used to quickly and meaningfully cohort patients using unsupervised classification methods. Additionally, to demonstrate increased translational value of machine learning models through the integration of nursing knowledge. Methods: A primary care practice dataset (N=3438) of high need patients defined by practice criteria was parsed to a subset population of patients with diabetes (n=1233). Three expert nurses selected variables for k-means cluster analysis using knowledge of critical factors for care coordination. Nursing knowledge was again applied to describe the psychosocial phenotypes in four prominent clusters, aligned with social and medical care plans. Results: Four distinct clusters interpreted and mapped to psychosocial need profiles, allowing for immediate translation to clinical practice through the creation of actionable social and medical care plans. (1) A large cluster of racially diverse female, non-English speakers with low medical complexity, and history of childhood illness; (2) A large cluster of English speakers with significant comorbidities (obesity and respiratory disease); (3) A small cluster of males with substance use disorder and significant comorbidities (mental health, liver and cardiovascular disease) who frequently visit the hospital; and (4) A moderate cluster of older, racially diverse patients with renal failure. Conclusions: This manuscript provides a practical method for analysis of primary care practice data using machine learning in tandem with expert clinical knowledge. Keywords: Social determinants of health; phenotypes; primary care; nursing; ambulatory care information systems; machine learning; care coordination; provider- provider communication; knowledge translation

A Systematic Review of the Application of Artificial Intelligence in Nursing Care: Where are We, and What’s Next?

“…Five included studies reveal that nurses implemented AI for patient risk identification. 17 , 21 , 23 , 26 , 27 For instance, a retrospective study by Brom et al utilized CART analysis and electronic health record (EHR) data to identify the risk of readmission in adult patients discharged from medical services. The results showed a 30-day readmission rate of 11.2%, and CART analysis revealed the highest risk for readmission among patients who visited the emergency department, had ≥ 9 comorbidities, were insured through Medicaid, and were ≥ 65 years old.…”

“…Four included studies demonstrate the utilization of AI by nurses to assist in developing nursing care plans. 18 , 20 , 21 , 27 In a discussion paper by Woodnutt et al, ChatGPT was employed to generate a mental health nursing care plan, and its output quality was evaluated against the authors’ clinical experience and existing guidance. The findings indicated that ChatGPT provided a care plan incorporating some principles of dialectical behavior therapy.…”

“…The integration of AI into nursing practices for risk identification presents a transformative opportunity to enhance patient outcomes. 17 , 21 , 23 , 26 , 27 The studies highlighted not only illustrate the diverse applications of AI and ML, encompassing the prediction of readmission risks and facilitating the management of chronic conditions but also emphasize the precision achieved through the identification of key risk factors. This precision enables targeted interventions, potentially leading to a reduction in healthcare costs and an overall improvement in patient care quality.…”

Medical-informed machine learning: integrating prior knowledge into medical decision systems

“…Embedding domain knowledge within ML models can add a layer of interpretability to the model [ 78 ] by providing the rationale behind the model’s recommendations [ 79 ] and allowing the re-traceability of the model decisions to specific model components [ 42 ]. Overall, these integration efforts also increase the transferability of the models and relevance in real-world medical settings [ 80 , 81 ], streamlining their integration into clinical practice. For these reasons, recent advancements in clinical AI focused on developing simpler, more interpretable models made of a minimal set of knowledge-driven features able to provide good explanations [ 50 , 73 ].…”

A clinical classification framework for identifying persons with high social and medical needs: The COMPLEXedex-SDH

“…Using the COMPLEXedex-SDH clinical algorithm, cases can be classified into actionable cohorts based on transparent metrics that the nurses can use to identify patients and prioritize interventions, such as outreach calls and cross-sector care alerts, to improve health equity and outcomes. While precision health techniques have been used in recent research to assist in the identification and classification of patients and utilization groups ( Hewner et al, 2023 ), the integration of value-based definitions attempts to capture individuals missed by traditional algorithms, improving reach to individuals disproportionally impacted by SDH.…”