Risk of Frequent Emergency Department Use Among an Ambulatory Care Sensitive Condition Population: A Population-based Cohort Study.

BACKGROUND: A small fraction of patients use a disproportionately large amount of emergency department (ED) resources. Identifying these patients, especially those with ambulatory care sensitive conditions (ACSC), would allow health care professionals to enhance their outpatient care. OBJECTIVE: The objectives of the study were to determine predictive factors associated with frequent ED use in a Quebec adult population with ACSCs and to compare several models predicting the risk of becoming an ED frequent user following an ED visit. RESEARCH DESIGN: This was an observational population-based cohort study extracted from Quebec's administrative data. SUBJECTS: The cohort included 451,775 adult patients, living in nonremote areas, with an ED visit between January 2012 and December 2013 (index visit), and previously diagnosed with an ACSC but not dementia. MEASURES: The outcome was frequent ED use (≥4 visits) during the year following the index visit. Predictors included sociodemographics, physical and mental comorbidities, and prior use of health services. We developed several logistic models (with different sets of predictors) on a derivation cohort (2012 cohort) and tested them on a validation cohort (2013 cohort). RESULTS: Frequent ED users represented 5% of the cohort and accounted for 36% of all ED visits. A simple 2-variable prediction model incorporating history of hospitalization and number of previous ED use accurately predicted future frequent ED use. The full model with all sets of predictors performed only slightly better than the simple model (area under the receiver-operating characteristic curve: 0.786 vs. 0.759, respectively; similar positive predictive value and number needed to evaluate curves). CONCLUSIONS: The ability to identify frequent ED users based only on previous ED and hospitalization use provides an opportunity to rapidly target this population for appropriate interventions.

Healthcare utilization after a first hospitalization for COPD: a new approach of State Sequence Analysis based on the '6W' multidimensional model of care trajectories.

BACKGROUND: Published methods to describe and visualize Care Trajectories (CTs) as patterns of healthcare use are very sparse, often incomplete, and not intuitive for non-experts. Our objectives are to propose a typology of CTs one year after a first hospitalization for Chronic Obstructive Pulmonary Disease (COPD), and describe CT types and compare patients' characteristics for each CT type. METHODS: This is an observational cohort study extracted from Quebec's medico-administrative data of patients aged 40 to 84 years hospitalized for COPD in 2013 (index date). The cohort included patients hospitalized for the first time over a 3-year period before the index date and who survived over the follow-up period. The CTs consisted of sequences of healthcare use (e.g. ED-hospital-home-GP-respiratory therapists, etc.) over a one-year period. The main variable was a CT typology, which was generated by a 'tailored' multidimensional State Sequence Analysis, based on the "6W" model of Care Trajectories. Three dimensions were considered: the care setting ("where"), the reason for consultation ("why"), and the speciality of care providers ("which"). Patients were grouped into specific CT types, which were compared in terms of care use attributes and patients' characteristics using the usual descriptive statistics. RESULTS: The 2581 patients were grouped into five distinct and homogeneous CT types: Type 1 (n = 1351, 52.3%) and Type 2 (n = 748, 29.0%) with low healthcare and moderate healthcare use respectively; Type 3 (n = 216, 8.4%) with high healthcare use, mainly for respiratory reasons, with the highest number of urgent in-hospital days, seen by pulmonologists and respiratory therapists at primary care settings; Type 4 (n = 100, 3.9%) with high healthcare use, mainly cardiovascular, high ED visits, and mostly seen by nurses in community-based primary care; Type 5 (n = 166, 6.4%) with high healthcare use, high ED visits and non-urgent hospitalisations, and with consultations at outpatient clinics and primary care settings, mainly for other reasons than respiratory or cardiovascular. Patients in the 3 highest utilization CT types were older, and had more comorbidities and more severe condition at index hospitalization. CONCLUSIONS: The proposed method allows for a better representation of the sequences of healthcare use in the real world, supporting data-driven decision making.

A comparative analysis of medication counting methods to assess polypharmacy in medico-administrative databases.

BACKGROUND: The variety of methods for counting medications may lead to confusion when attempting to compare the extent of polypharmacy across different populations. OBJECTIVE: To compare the prevalence estimates of polypharmacy derived from medico-administrative databases, using different methods for counting medications. METHODS: Data were drawn from the Québec Integrated Chronic Disease Surveillance System. A random sample of 110,000 individuals aged >65 was selected, including only those who were alive and covered by the public drug plan during the one-year follow-up. We used six methods to count medications: #1-cumulative one-year count, #2-average of four quarters' cumulative counts, #3-count on a single day, #4-count of medications used in first and fourth quarters, #5-count weighted by duration of exposure, and #6-count of uninterrupted medication use. Polypharmacy was defined as ≥5 medications. Cohen's Kappa was calculated to assess the level of agreement between the methods. RESULTS: A total of 93,516 (85 %) individuals were included. The prevalence of polypharmacy varied across methods. The highest prevalence was observed with cumulative methods (#1:74.1 %; #2:61.4 %). Single day count (#3:47.6 %), first and fourth quarters count (#4:49.5 %), and weighted count (#5:46.6 %) yielded similar results. The uninterrupted use count yielded the lowest estimate (#6:35.4 %). The weighted method (#5) showed strong agreement with the first and fourth quarters count (#4). Cumulative methods identified higher proportions of younger, less multimorbid individuals compared to other methods. CONCLUSION: Counting methods significantly affect polypharmacy prevalence estimates, necessitating their consideration when comparing and interpretating results.

Machine learning to improve frequent emergency department use prediction: a retrospective cohort study.

Frequent emergency department use is associated with many adverse events, such as increased risk for hospitalization and mortality. Frequent users have complex needs and associated factors are commonly evaluated using logistic regression. However, other machine learning models, especially those exploiting the potential of large databases, have been less explored. This study aims at comparing the performance of logistic regression to four machine learning models for predicting frequent emergency department use in an adult population with chronic diseases, in the province of Quebec (Canada). This is a retrospective population-based study using medical and administrative databases from the Régie de l'assurance maladie du Québec. Two definitions were used for frequent emergency department use (outcome to predict): having at least three and five visits during a year period. Independent variables included sociodemographic characteristics, healthcare service use, and chronic diseases. We compared the performance of logistic regression with gradient boosting machine, naïve Bayes, neural networks, and random forests (binary and continuous outcome) using Area under the ROC curve, sensibility, specificity, positive predictive value, and negative predictive value. Out of 451,775 ED users, 43,151 (9.5%) and 13,676 (3.0%) were frequent users with at least three and five visits per year, respectively. Random forests with a binary outcome had the lowest performances (ROC curve: 53.8 [95% confidence interval 53.5-54.0] and 51.4 [95% confidence interval 51.1-51.8] for frequent users 3 and 5, respectively) while the other models had superior and overall similar performance. The most important variable in prediction was the number of emergency department visits in the previous year. No model outperformed the others. Innovations in algorithms may slightly refine current predictions, but access to other variables may be more helpful in the case of frequent emergency department use prediction.