Patient Journey Modelling – from data to insights into patient pathways
Flagship Program: Intelligent Decision Support to Improve Value and Efficiency
NSW Health is responding to challenges facing Australian health systems with several state-wide priority programs focused on the core strategic ambition to deliver value-based healthcare. With average age increasing and a growing number of people living with chronic or complex health conditions, the NSW population health needs are changing and demands on the health system are rising. NSW aspires to provide outcome-focused and value-based care: improving quality of care, and the health and experience of the population and providers whist reducing cost.
Key strategic initiatives of NSW Health aim to transform the health system to routinely deliver person-centred, seamless, efficient and effective care, particularly for people with complex, long term conditions. The focus is on delivering evidence based, coordinated, patient-centred and outcome-focused care in the community. These include Leading Better Value Care, Integrated Care and, most recently, Collaborative Commissioning.
Collaborative Commissioning (CC) is part of the strategic direction of NSW to deliver value-based healthcare. There are significant opportunities to deliver value-based health care by local partnerships focusing on care in the community. CC is a whole-of-system approach designed to enable and support delivery of value-based care in the community. It aims to incentivise locally developed integration of care across the entire continuum and embed local accountability for delivering value-driven, outcome focused and patient-centred care. To achieve the desired outcomes this requires collaboration between LHDs, PHNs and other service providers to support people’s health needs in the community, without the need for hospital-based care when it could be avoided.
Collaborative Commissioning commenced in 2019 with the initiation of a joint development phase of PHN/LHD CC partnerships. The Western Sydney Partnership (including Western Sydney PHN and LHD) have proposed two models of care under CC. These are being developed through their Joint Development Phase (JDP) with support from the NSW Ministry of Health (MoH). These models encourage innovative strategic planning to re-design pathways for locally defined priority cohorts by leveraging already existing local resources. Western Sydney are developing their pathways for value based urgent care and cardiology in the community.
Effective CC requires insightful understanding of past and future patient pathways. The design, implementation and evaluation of collaboratively commissioned care pathways across the entire continuum of care, spanning the acute setting and in the community setting, is envisaged to rest on insights from an evidence base. The needed insights include the following.
- What determines health care choices by patients and clinicians?
- Where are crucial transitions in care where important health care choices are made?
- What are the patterns of access to health care for cohorts of interest?
The MoH has multiple initiatives to address the need for insight. Initiatives include collaboration with the data analytics and Lumos teams in MoH Systems Information and Branch Analytics Branch, establishing the Dynamic Simulation Modelling (DSM) project, and the Patient Journey Modelling project described in this PCP.
To date, understanding patient pathways has been impeded by lack of data to provide a comprehensive picture of the patient pathway across primary, acute and non-admitted health settings. This data gap has recently been addressed by the NSW Health Lumos data asset. This initiative links data across primary, ambulatory and acute care establishing Australia’s first linked Primary Care Data Asset.
The Lumos data asset can support the integration and reorientation of the NSW health system through Collaborative Commissioning with appropriate data. Leveraging Lumos, the present project proposes to develop a new approach to delivering essential data informed insights about patient pathways to inform Collaborative Commissioning, A key challenge is how to make sense of large numbers of complex journeys and how to use massive patient journey data to provide decision support technology. An individual patient’s journey is a collection of time stamped records, where each record describes the patient’s interaction with the health system. We now amass much data on patients’ journeys. An example is the Lumos data asset containing de-identified unit records for over a million patients across NSW, covering events such as GP visits, ED presentation and hospital admissions.
This project proposes to take on the challenge of enabling CC partnerships to harness the value of Lumos. A common way to extract information from patient journeys is to build a predictive model that quantifies the chance of some future event. Most models are focused through a combination of (1) the event of interest, (2) the future time horizon of interest, (3) patient cohort of interest. Different models may also use different approaches for quantifying chance. Another approach uses predictive models to dynamically simulate individual patient pathways as they unfold in time. This is a way to extrapolate system demand (and interaction with supply) as it allows for ultimate flexibility in the analysis of cohorts of patients, e.g., for understanding shortfalls, costs and outcomes of health service delivery. Simulation offers a significant improvement over collections of predictive models. Both simulation, and datasets of real patient pathways provide many data points of patients and their events. A key question is, how to analyse these many data points to deliver insights. To date such analysis is limited to providing key statistics, trends, correlations and plots.
The primary aim of this project is to develop a methodology for modelling of patient pathways from patient journeys. A patient journey is the collection of health events of an individual patient. A patient pathway is a sequence (or network) of health events types that abstracts and groups patient journeys. The methodology developed by this project aims to be (i) robust to the different ways of conceptualizing patient journeys and pathways, (ii) rigorous in that it is based on accepted formalisms and stands up to peer review, and (iii) reusable across different diseases and models of care.
A secondary aim of this project is to apply the methodology to selected Collaborative Commissioning initiatives to provide local insights to support Collaborative Commissioning. The project aims to provide the detailed understanding of the individual patient journeys of a cohort within a population, for example local cohorts in the case of CC. These insights, knowing why things happen through understanding typical local patient journeys, will provide an evidence base for CC to develop pathways with the most benefit for the patient and local need. The ambition is to support strategic decisions for CC pathway design.
Additional aims of this project are to realize the developed methodology as proof-of-concept software. This will support the aim of applying the methodology and the aim for reusability. An aspect of this aim is to realize insights using data visualization, which may involve the development of visualization tools.
This project will work in parallel with and complement other Collaborative Commissioning initiatives. This project will work closely with the Systems Information Monitoring and Evaluation team in MoH, and with the Dynamic Simulation Modelling (DSM) project run by the Collaborative Commissioning team.
This project will assist NSW Health and Collaborative Commissioning Partnerships to better understand patients’ interactions with the health system and how the characteristics of these patients and their prototypical journeys can inform strategic decisions. This project will not include provisions to establish new datasets, nor to create user interfaces or dashboards for use by clinicians or managers. Neither is there provision for integration into existing decision support platforms.