“Bow-tie” optimal pathway discovery analysis of sepsis hospital admissions using the Hospital Episode Statistics database in England

Abstract

Objective The “Bow-tie” optimal pathway discovery analysis uses large clinical event datasets to map clinical pathways and to visualize risks (improvement opportunities) before, and outcomes after, a specific clinical event. This proof-of-concept study assesses the use of NHS Hospital Episode Statistics (HES) in England as a potential clinical event dataset for this pathway discovery analysis approach. Materials and Methods A metaheuristic optimization algorithm was used to perform the “bow-tie” analysis on HES event log data for sepsis (ICD-10 A40/A41) in 2016. Analysis of hospital episodes across inpatient and outpatient departments was performed for the period 730 days before and 365 days after the index sepsis hospitalization event. Results HES data captured a sepsis event for 76 523 individuals (>13 years), relating to 580 000 coded events (across 220 sepsis and non-sepsis event classes). The “bow-tie” analysis identified several diagnoses that most frequently preceded hospitalization for sepsis, in line with the expectation that sepsis most frequently occurs in vulnerable populations. A diagnosis of pneumonia (5 290 patients) and urinary tract infections (UTIs; 2 057 patients) most often preceded the sepsis event, with recurrent UTIs acting as a potential indicative risk factor for sepsis. Discussion This proof-of-concept study demonstrates that a “bow-tie” pathway discovery analysis of the HES database can be undertaken and provides clinical insights that, with further study, could help improve the identification and management of sepsis. The algorithm can now be more widely applied to HES data to undertake targeted clinical pathway analysis across multiple healthcare conditions.

Process Mining of Disease Trajectories: A Literature Review

“…Table 1 summarizes the four articles identified at the conclusion of our literature search [9][10][11][12]. The four papers use four different data sources, four different model visualizations, three different trajectory approaches and three different conformance checking metrics.…”

“…The four papers use four different data sources, four different model visualizations, three different trajectory approaches and three different conformance checking metrics. [11]; (4) a synthetic EHR dataset used to demonstrate the concept of process mining to identify disease trajectories (n=50) [12]. All articles used the International Classification of Disease (ICD) codes published by the World Health Organisation (WHO) [14], either the 9th revision (ICD-9) [9,10] or the 10th revision (ICD-10) which was used by [11] and [12].…”

“…Patient selection based on a specific diagnosis was performed by de Toledo et al [10] with Type-2 Diabetes and by De Oliveira et al [11] with a focus on patients with sepsis. In contrast to these, Kusuma et al [9] did not pre-select diseases of interest but did excluded those ICD codes that do not relate directly to a diagnosis.…”

“…Three approaches were identified to define the order of the diagnostic codes to form a trajectory: (1) a set of statistical analysis of correlation measurement and a binomial test [9]; (2) n-grams [10]; (3) metaheuristics optimisation algorithm [11]. The implementation of process mining in the study of disease trajectories was found in three ways: process discovery to build the disease trajectory model, conformance checking, and for model visualisation.…”

“…The Fuzzy Miner and Heuristics Miner algorithms [15] were used by de Toledo et al [10] for mining the event traces and for visualise the model respectively. A proprietary Metaheuristics optimisation algorithm [16] was used by De Oliveira et al [11] while a private-company-developed app was used to visualise the model. Kusuma et al [9] used the interactive Data-aware Heuristics Miner [15] to mine the model, and the directly-followed graph for model visualisation.…”