Abstract
Purpose of the paper: Health economists have set the challenge for themselves to not only act in their traditional “technology evaluator” role but also facilitate a greater “search for efficiency” approach within health systems (Scotland & Bryan, 2017). This involves the iterative management of existing care, along integrated care pathways for whole conditions, rather than the additive adoption of discrete and disjointed products and services. But this is no simple task. It necessitates the involvement of value-based thinking much earlier within service planning and intervention design. It also requires participatory methods to engage diverse clinical, administrative and executive staff to understand the expected value of their different options (Partington & Karnon, 2021). The deliberation and interpretation of local utilisation and outcomes data is key to ensuring that evidence is directly generalisable to idiosyncratic, dynamic and adaptive decision-problems (Braithwaite, 2018). We have been trialling methods for early and exploratory economic evaluation, while embedded within Local Health Networks in South Australia.
Methods: We linked routinely collected costing, administrative and clinical data to describe patient pathways through the hospital system. Data were structured and analysed within a decision-analytic framework, using process mining and decision-trees, to represent the observed probability of different pathways and magnitude of resulting costs, operational indicators, and health outcomes.
Case presentation: Existing care was described for patients aged 65+ years old who present to the Emergency Department (ED) via ambulance and receive a triage category of 3 to 5. Inputs and outcomes of interest included the distributions in length-of-stay (LOS), broken-down by steps in the pathways; distributions in pathway costs, broken down by cost-buckets; the associated activity-based funding and the resulting ‘financial impact’; and rates of representations, readmissions, and mortality.
Summary of results: Existing care was profiled for 5,400 patients who presented in the first half of 2020. The decision-tree visualises the relative proportion of patients who traverse different pathways through non-admitted, acute-admitted and sub/non-acute admitted care. While the average ED and inpatient LOS for admitted patients was 5.4 hours and 4.1 days, respectively, significant variation is seen within and across pathways. Some pathways lead to positive financial impacts, while others do not.
In viewing the results, the LHN expect that a proportion of patients who traverse particular pathways can be redirected along alternative routes or pass along the same pathway, more efficiency. Using the observational data to inform a structured elicitation exercise we show how these expectations, and their associated uncertainty and disagreement, are quantified. Elicited parameters are fed-back into the model alongside estimates of required resources, so that these expectations can be tested and updated.
Conclusions: We show how local data-driven and elicitation-informed decision-analytic modelling can be used in formative evaluations of potential interventions, prior to their implementation and as part of a continuous improvement process. The approach is translatable to any setting with similar data collections and stakeholders. Such modelling might help LHNs develop ‘reference models’ around segments of their patient populations for whom to then iteratively explore the value of their healthcare provision.
Methods: We linked routinely collected costing, administrative and clinical data to describe patient pathways through the hospital system. Data were structured and analysed within a decision-analytic framework, using process mining and decision-trees, to represent the observed probability of different pathways and magnitude of resulting costs, operational indicators, and health outcomes.
Case presentation: Existing care was described for patients aged 65+ years old who present to the Emergency Department (ED) via ambulance and receive a triage category of 3 to 5. Inputs and outcomes of interest included the distributions in length-of-stay (LOS), broken-down by steps in the pathways; distributions in pathway costs, broken down by cost-buckets; the associated activity-based funding and the resulting ‘financial impact’; and rates of representations, readmissions, and mortality.
Summary of results: Existing care was profiled for 5,400 patients who presented in the first half of 2020. The decision-tree visualises the relative proportion of patients who traverse different pathways through non-admitted, acute-admitted and sub/non-acute admitted care. While the average ED and inpatient LOS for admitted patients was 5.4 hours and 4.1 days, respectively, significant variation is seen within and across pathways. Some pathways lead to positive financial impacts, while others do not.
In viewing the results, the LHN expect that a proportion of patients who traverse particular pathways can be redirected along alternative routes or pass along the same pathway, more efficiency. Using the observational data to inform a structured elicitation exercise we show how these expectations, and their associated uncertainty and disagreement, are quantified. Elicited parameters are fed-back into the model alongside estimates of required resources, so that these expectations can be tested and updated.
Conclusions: We show how local data-driven and elicitation-informed decision-analytic modelling can be used in formative evaluations of potential interventions, prior to their implementation and as part of a continuous improvement process. The approach is translatable to any setting with similar data collections and stakeholders. Such modelling might help LHNs develop ‘reference models’ around segments of their patient populations for whom to then iteratively explore the value of their healthcare provision.
Original language | English |
---|---|
Publication status | Published - 6 May 2022 |
Event | Independent Health and Aged Care Pricing Authority Activity Based Funding Conference 2022: Innovation and collaboration: Activity based funding for sustainability in health care - Brisbane, Australia Duration: 5 May 2022 → 6 May 2022 Conference number: 9th https://www.abfconference.com.au/ |
Conference
Conference | Independent Health and Aged Care Pricing Authority Activity Based Funding Conference 2022 |
---|---|
Abbreviated title | IHACPA 2022 |
Country/Territory | Australia |
City | Brisbane |
Period | 5/05/22 → 6/05/22 |
Other | The Activity Based Funding Conference series is an opportunity for professional development, knowledge-sharing and public discourse on the innovative work taking place in the health care sector through the application of activity based funding. The Activity Based Funding Conference 2022 event was relevant to professionals with an interest in learning more about the activity based funding work taking place in Australia, as well as discovering international perspectives. |
Internet address |
Keywords
- Health technology
- Health economics
- Health services
- Integrated health care
- Health system sustainability