Children with Medical Complexity: Evaluating Reported Costs and Modelling Predictors of Hospital and ED Resource Use

• Author / Creator

  Sidra, Michael S.

• Children with medical complexity (CMCs) are a vulnerable population that suffers from complex chronic conditions that require intensive health care resources. Previous studies, largely focused on hospitalizations and emergency department visits, have examined how clinical factors are associated with health care resource use for this population. However, the impacts of socioeconomic factors are less studied. There is also little information on the Alberta CMC population, beyond broad quantifications of health care resource use at the national level. The objective of this research was first to identify how health care costs for CMCs are reported in the literature and then to examine how both clinical and socioeconomic factors are jointly associated with hospital days and emergency department use after an index admission (first admission with a diagnosis of Complex Chronic Condition) among CMCs in Alberta, and how these associations change over time.

  This dissertation first presents a systematic review outlining how CMCs are identified in administrative health data together with the types and methods of cost reporting for CMCs. The review identifies limitations in interpreting costing trends and summarizes recommendations from the literature on CMC health care resource costing.

  The dissertation then studies a population-based cohort of CMCs in Alberta through separate inferential and predictive analyses to quantitatively assess how clinical and socioeconomic factors are associated with hospitalizations and ED visits over time. The first analysis utilizes a mixed-effect linear hurdle regression model to longitudinally estimate and perform inference for these associations. The second analysis uses a tree-based gradient-boosted regression model to predict and identify the most important predictors of resource use in the short and long terms (i.e., 1 and 5 years after index admission, respectively). The analysis leads into a discussion of the opportunities and limitations of using administrative health data and electronic health records in predictive machine learning models in CMC-related research.

  The first analysis found that initial length of stay (LOS) and number of chronic medications, both proxies for clinical complexity, have strong, positive associations with resource use. Specifically, a greater initial LOS was significantly, positively associated with more hospital days whereas more chronic medications were significantly, positively associated with more hospital days and more ED visits in the first year after initial discharge. In terms of socioeconomic factors, the analysis found that CMCs living in rural and remote rural areas had more ED visits than those living in urban/metropolitan locations. Material and social deprivation, here measured with Canadian census data, also had significant, positive associations with ED visits.

  The second analysis also showed that clinical proxy measures namely initial LOS and clinical classification (single vs. multiple complex chronic conditions) were top predictors of hospital days. The top predictors of ED use were consistently socioeconomic in nature, with patient residence being an important predictor. The results highlighted the need for more-detailed clinical and socioeconomic data in electronic health records and supported further discussion of how predictive modelling can be used in the future to enhance understanding of the CMC population.

  Overall, that both analyses pointed to the relationships between clinical complexity and hospital use and between socioeconomic status and ED visits is an important contribution to the CMC literature and, more specifically, to policy development efforts and health care administrators aiming to improve care for the Alberta CMC population. Further research on the availability of health care resources for CMCs living in (remote) rural areas of Alberta is warranted. Finally, this dissertation demonstrated that predictive and inferential modelling using administrative health data can be used to identify factors associated with health care use by CMCs.

• Subjects / Keywords

  • Children with Medical Complexity
  • Hospital Use by Children with Medical Complexity
  • Emergency Department Use by Children with Medical Complexity

• Graduation date

  Fall 2023

• Type of Item

  Thesis

• Degree

  Doctor of Philosophy

• DOI

  https://doi.org/10.7939/r3-nk9a-3b76

• License

  This thesis is made available by the University of Alberta Libraries with permission of the copyright owner solely for non-commercial purposes. This thesis, or any portion thereof, may not otherwise be copied or reproduced without the written consent of the copyright owner, except to the extent permitted by Canadian copyright law.