Extracting, analyzing and using patterns of service utilization from a cross continuum health service system to optimize care for patients with complex issues – a methodological approach.
| dc.contributor.author | Bambi, Jonas | |
| dc.contributor.supervisor | Kuo, Alex | |
| dc.date.accessioned | 2024-09-09T19:24:13Z | |
| dc.date.available | 2024-09-09T19:24:13Z | |
| dc.date.issued | 2024 | |
| dc.degree.department | School of Health Information Science | |
| dc.degree.level | Doctor of Philosophy PhD | |
| dc.description.abstract | Introduction: The clinical service system performs a very large and diverse array of functions associated with a broadly differentiated array of problems. As a result, the service system is broken up functionally and administratively into service units. To provide the best possible care for patients with chronic or complex problems the interoperation of service system components needs to be optimized. Such an optimization requires the enactment of appropriate problem-specific clinical protocols as well as cohort-specific Clinical Practice Guidelines (CPGs). However, alignment of the service system operations with CPGs is very challenging. This is mostly due to the challenge of identifying longitudinal patterns of service utilization (PSUs) in a cross-continuum data to assess adherence to the CPGs. Hence, the research activities aim to address the following questions: (1) what methodology can be employed to extract and identify clinically understandable cohorts-specific patients’ PSUs within sparse high-dimensional cross-continuum healthcare datasets? (2) to what extent can one cut across the complexity of a cross-continuum service structure to capture the dynamics of the journey of patients with complex issues that clearly portray their engagement with the service system, to help locate potential operational problems? (3) once identified, to what extent can PSUs be used to inform quality assurance and quality improvement (QA/QI) initiatives? Methods: Starting with a semantic layer, referred to as the Clinical Context Coding Scheme, to address data granularity and nomenclature issues, various machine learning approaches were used to extract PSUs. These include the use of nested-iterative graph community detection, directed graph, and Natural Language Processing (NLP) clustering. These steps were followed by the use of various graph metrics and input from clinical and operational subject matter experts to refine the level of resolution for the extracted patterns. Results: The results have shown that by using a nested-iterative community detection or NLP clustering, it is possible to extract cohorts-specific high-prevalence functionally integrated PSUs. The results have also shown that directed graphs are well suited to the task of depicting the way that the diverse components of the system are functionally coupled—or remain disconnected—by patients journeys. Discussion: These findings have several implications related to the optimization of care for patients with chronic/complex problems, including: (1) the possibility of influencing the reorganization of some services within a health organization service-structure to provide the most optimal connections between services to address patients’ needs, and (2) the first step in addressing the challenge of locating potential operational problems for patients with complex issues engaging with a complex healthcare service system. Additionally, the combination of the proposed methodologies with various statistical analysis, have demonstrated that PSUs can play an important role in informing diverse QA/QI initiatives, including: (1) providing a nuanced approach to assess and measure access disparity, based on local reality, across the full care continuum, and a first step in addressing inequities for a healthcare organization, and (2) equipping a healthcare organization with required information, based on local reality, to provide better care for the opioid overdose patients, as well as being pro-active in preventing subsequent overdoses. However, in informing QA/QI initiatives, distal factors such as social determinants of health not captured within the dataset are not considered in the models, limiting the usefulness of the recommendations. This is a limitation that future research will need to address. Conclusion: The research activities undertaken provide a first step – the extraction of PSUs, in introducing a novel analytics framework relying on patients’ service pathways as a foundation to evaluate conformance of interventions to cohort-specific CPGs. In collaboration with various clinical and operational SMEs, future research will expand on PSUs, to include other elements required for this novel analytical framework to be used as one of the paradigms to the secondary use of data collected by health organizations, to support the implementation of Learning Healthcare Systems. | |
| dc.description.scholarlevel | Graduate | |
| dc.identifier.bibliographicCitation | Bambi, J., Santoso, Y., Sadri, H., Moselle, K., Rudnick, A., Robertson, S., ... & Richardson, A. (2024). A Methodological Approach to Extracting Patterns of Service Utilization from a Cross-Continuum High Dimensional Healthcare Dataset to Support Care Delivery Optimization for Patients with Complex Problems. BioMedInformatics, 4(2), 946-965. | |
| dc.identifier.bibliographicCitation | Bambi, J., Sadri, H., Moselle, K., Chang, E., Santoso, Y., Howie, J., ... & Kuo, A. (2024). Approaches to Extracting Patterns of Service Utilization for Patients with Complex Conditions: Graph Community Detection vs. Natural Language Processing Clustering. BioMedInformatics, 4(3), 1884-1900. | |
| dc.identifier.bibliographicCitation | Bambi, J., Santoso, Y., Moselle, K., Robertson, S., Rudnick, A., Chang, E., & Kuo, A. (2024). Analyzing Patterns of Service Utilization Using Graph Topology to Understand the Dynamic of the Engagement of Patients with Complex Problems with Health Services. BioMedInformatics, 4(2), 1071-1084. | |
| dc.identifier.bibliographicCitation | Bambi, J., Dong, G. Y., Santoso, Y., Moselle, K., Dugas, S., Olobatuyi, K., ... & Kuo, A. (2024). Patterns of Service Utilization across the Full Continuum of Care: Using Patient Journeys to Assess Disparities in Access to Health Services. Knowledge, 4(2), 252-264. | |
| dc.identifier.bibliographicCitation | Bambi, J., Olobatuyi, K., Santoso, Y., Sadri, H., Moselle, K., Rudnick, A., ... & Kuo, A. (2024). Use of Patterns of Service Utilization and Hierarchical Survival Analysis in Planning and Providing Care for Overdose Patients and Predicting the Time-to-Second Overdose. Knowledge, 4(3), 444-461. | |
| dc.identifier.uri | https://hdl.handle.net/1828/20392 | |
| dc.language | English | eng |
| dc.language.iso | en_US | |
| dc.rights | Available to the World Wide Web | |
| dc.subject | clinical practice guidelines | |
| dc.subject | clinical pathways | |
| dc.subject | decision support | |
| dc.subject | graph community detection | |
| dc.subject | health information management | |
| dc.subject | health service system | |
| dc.subject | machine learning algorithms | |
| dc.subject | clustering algorithms | |
| dc.subject | electronic healthcare | |
| dc.subject | natural language processing | |
| dc.subject | graph topology | |
| dc.subject | patients’ engagement dynamics | |
| dc.subject | public healthcare | |
| dc.subject | mental health | |
| dc.subject | healthcare equity | |
| dc.subject | healthcare access | |
| dc.subject | access disparities | |
| dc.subject | opioid overdose | |
| dc.subject | opioid crisis | |
| dc.subject | survival analysis | |
| dc.title | Extracting, analyzing and using patterns of service utilization from a cross continuum health service system to optimize care for patients with complex issues – a methodological approach. | |
| dc.type | Thesis |