Process mining for healthcare decision analytics with micro-costing estimations.

Managing constrained healthcare resources is an important and inescapable role of healthcare decision makers. Allocative decisions are based on downstream consequences of changes to care processes: judging whether the costs involved are offset by the magnitude of the consequences, and therefore whether the change represents value for money. Process mining techniques can inform such decisions by quantitatively discovering, comparing and detailing care processes using recorded data, however the scope of techniques typically excludes anything 'after-the-process' i.e., their accumulated costs and resulting consequences. Cost considerations are increasingly incorporated into process mining techniques, but the majority of healthcare costs for service and overhead components are commonly apportioned and recorded at the patient (trace) level, hiding event level detail. Within decision-analysis, event-driven and individual-level simulation models are sometimes used to forecast the expected downstream consequences of process changes, but are expensive to manually operationalise. In this paper, we address both of these gaps within and between process mining and decision analytics, by better linking them together. In particular, we introduce a new type of process model containing trace data that can be used in individual-level or cohort-level decision-analytical model building. Furthermore, we enhance these models with process-based micro-costing estimations. The approach was evaluated with health economics and decision modelling experts, with discussion centred on how the outputs could be used, and how similar information would otherwise be compiled.

Three novel bird strike likelihood modelling techniques: The case of Brisbane Airport, Australia.

The risk posed by wildlife to air transportation is of great concern worldwide. In Australia alone, 17,336 bird-strike incidents and 401 animal-strike incidents were reported to the Air Transport Safety Board (ATSB) in the period 2010-2019. Moreover, when collisions do occur, the impact can be catastrophic (loss of life, loss of aircraft) and involve significant cost to the affected airline and airport operator (estimated at globally US$1.2 billion per year). On the other side of the coin, civil aviation, and airport operations have significantly affected bird populations. There has been an increasing number of bird strikes, generally fatal to individual birds involved, reported worldwide (annual average of 12,219 reported strikes between 2008-2015 being nearly double the annual average of 6,702 strikes reported 2001-2007) (ICAO, 2018). Airport operations including construction of airport infrastructure, frequent take-offs and landings, airport noise and lights, and wildlife hazard management practices aimed at reducing risk of birdstrike, e.g., spraying to remove weeds and invertebrates, drainage, and even direct killing of individual hazard species, may result in habitat fragmentation, population decline, and rare bird extinction adjacent to airports (Kelly T, 2006; Zhao B, 2019; Steele WK, 2021). Nevertheless, there remains an imperative to continually improve wildlife hazard management methods and strategies so as to reduce the risk to aircraft and to bird populations. Current approved wildlife risk assessment techniques in Australia are limited to ranking of identified hazard species, i.e., are 'static' and, as such, do not provide a day-to-day risk/collision likelihood. The purpose of this study is to move towards a dynamic, evidence-based risk assessment model of wildlife hazards at airports. Ideally, such a model should be sufficiently sensitive and responsive to changing environmental conditions to be able to inform both short and longer term risk mitigation decisions. Challenges include the identification and quantification of contributory risk factors, and the selection and configuration of modelling technique(s) that meet the aforementioned requirements. In this article we focus on likelihood of bird strike and introduce three distinct, but complementary, assessment techniques, i.e., Algebraic, Bayesian, and Clustering (ABC) for measuring the likelihood of bird strike in the face of constantly changing environmental conditions. The ABC techniques are evaluated using environment and wildlife observations routinely collected by the Brisbane Airport Corporation (BAC) wildlife hazard management team. Results indicate that each of the techniques meet the requirements of providing dynamic, realistic collision risks in the face of changing environmental conditions.

xPM: Enhancing exogenous data visibility.

Process mining is a well-established discipline with applications in many industry sectors, including healthcare. To date, few publications have considered the context in which processes execute. Little consideration has been given as to how contextual data (exogenous data) can be practically included for process mining analysis, beyond including case or event attributes in a typical event log. We show that the combination of process data (endogenous) and exogenous data can generate insights not possible with standard process mining techniques. Our contributions are a framework for process mining with exogenous data and new analyses, where exogenous data and process behaviour are linked to process outcomes. Our new analyses visualise exogenous data, highlighting the trends and variations, to show where overlaps or distinctions exist between outcomes. We applied our analyses in a healthcare setting and show that clinicians could extract insights about differences in patients' vital signs (exogenous data) relevant to clinical outcomes. We present two evaluations, using a publicly available data set, MIMIC-III, to demonstrate the applicability of our analysis. These evaluations show that process mining can integrate large amounts of physiologic data and interventions, with resulting discrimination and conversion to clinically interpretable information.

Trauma by-pass guideline: A data-driven conformance analysis for road trauma cases in Queensland.

OBJECTIVE: Study objectives were to (i) develop and test a whole-of-system method for identifying patients who meet a major trauma by-pass guideline definition; (ii) apply this method to assess conformance to the current 2006 guideline for a road trauma cohort; and (iii) leverage relevant findings to propose improvements to the guideline. METHODS: Retrospective analysis of existing, routinely collected data relating to Queensland road trauma patients July 2015 to June 2017. Data from ambulance, aero-medical retrievals, ED, hospital and death registers were linked and used for analysis. Processes of care measured included: frequency of pre-hospital triage criteria, distribution of destination (trauma service level), compliance with guideline (recommended vs actual destination), trauma service level by threat to life (injury severity) (all modes of transport and aero-medical in particular), proportion of patients requiring only ED, transport pathway (direct vs inter-hospital transfer). RESULTS: 3847 cases were identified from data as meeting criteria for major trauma by-pass. The top five most frequently used criteria for qualifying patients as meeting the major trauma by-pass guideline were pulse rate, vehicle rollover, possible spinal cord injury, respiration rate and entrapment. The study demonstrates a 65% conformance to the clinical guideline. Overtriaged patients (transported to higher trauma service than recommended) generally reveal International Classification of Disease Injury Severity Score representing a high threat to life. CONCLUSION: Overall, the present study found good conformance, with overtriage rate as expected by clinicians. It is recommended to include data values to capture paramedics assessment of trauma level to enable more accurate assessment of conformance to guideline and future revision of the thresholds.

A Comparative Process Mining Analysis of Road Trauma Patient Pathways.

In this paper we report on key findings and lessons from a process mining case study conducted to analyse transport pathways discovered across the time-critical phase of pre-hospital care for persons involved in road traffic crashes in Queensland (Australia). In this study, a case is defined as being an individual patient's journey from roadside to definitive care. We describe challenges in constructing an event log from source data provided by emergency services and hospitals, including record linkage (no standard patient identifier), and constructing a unified view of response, retrieval, transport and pre-hospital care from interleaving processes of the individual service providers. We analyse three separate cohorts of patients according to their degree of interaction with Queensland Health's hospital system (C1:no transport required, C2:transported but no Queensland Health hospital, C3:transported and hospitalisation). Variant analysis and subsequent process modelling show high levels of variance in each cohort resulting from a combination of data collection, data linkage and actual differences in process execution. For Cohort 3, automated process modelling generated 'spaghetti' models. Expert-guided editing resulted in readable models with acceptable fitness, which were used for process analysis. We also conduct a comparative performance analysis of transport segment based on hospital `remoteness'. With regard to the field of process mining, we reach various conclusions including (i) in a complex domain, the current crop of automated process algorithms do not generate readable models, however, (ii) such models provide a starting point for expert-guided editing of models (where the tool allows) which can yield models that have acceptable quality and are readable by domain experts, (iii) process improvement opportunities were largely suggested by domain experts (after reviewing analysis results) rather than being directly derived by process mining tools, meaning that the field needs to become more prescriptive (automated derivation of improvement opportunities).

Privacy-Preserving Process Mining in Healthcare.

Process mining has been successfully applied in the healthcare domain and has helped touncover various insights for improving healthcare processes. While the benefits of process miningare widely acknowledged, many people rightfully have concerns about irresponsible uses of personaldata. Healthcare information systems contain highly sensitive information and healthcare regulationsoften require protection of data privacy. The need to comply with strict privacy requirements mayresult in a decreased data utility for analysis. Until recently, data privacy issues did not get muchattention in the process mining community; however, several privacy-preserving data transformationtechniques have been proposed in the data mining community. Many similarities between datamining and process mining exist, but there are key differences that make privacy-preserving datamining techniques unsuitable to anonymise process data (without adaptations). In this article, weanalyse data privacy and utility requirements for healthcare process data and assess the suitabilityof privacy-preserving data transformation methods to anonymise healthcare data. We demonstratehow some of these anonymisation methods affect various process mining results using three publiclyavailable healthcare event logs. We describe a framework for privacy-preserving process mining thatcan support healthcare process mining analyses. We also advocate the recording of privacy metadatato capture information about privacy-preserving transformations performed on an event log.

Leveraging Data Quality to Better Prepare for Process Mining: An Approach Illustrated Through Analysing Road Trauma Pre-Hospital Retrieval and Transport Processes in Queensland.

While noting the importance of data quality, existing process mining methodologies (i) do not provide details on how to assess the quality of event data (ii) do not consider how the identification of data quality issues can be exploited in the planning, data extraction and log building phases of any process mining analysis, (iii) do not highlight potential impacts of poor quality data on different types of process analyses. As our key contribution, we develop a process-centric, data quality-driven approach to preparing for a process mining analysis which can be applied to any existing process mining methodology. Our approach, adapted from elements of the well known CRISP-DM data mining methodology, includes conceptual data modeling, quality assessment at both attribute and event level, and trial discovery and conformance to develop understanding of system processes and data properties to inform data extraction. We illustrate our approach in a case study involving the Queensland Ambulance Service (QAS) and Retrieval Services Queensland (RSQ). We describe the detailed preparation for a process mining analysis of retrieval and transport processes (ground and aero-medical) for road-trauma patients in Queensland. Sample datasets obtained from QAS and RSQ are utilised to show how quality metrics, data models and exploratory process mining analyses can be used to (i) identify data quality issues, (ii) anticipate and explain certain observable features in process mining analyses, (iii) distinguish between systemic and occasional quality issues, and (iv) reason about the mechanisms by which identified quality issues may have arisen in the event log. We contend that this knowledge can be used to guide the data extraction and pre-processing stages of a process mining case study to properly align the data with the case study research questions.