Design & Implementation of a Learning Health System in Australia Data Dissect Pty Ltd (datadissect.com.au) Tom Cundy, Stefan Court-Kowalski, Andrew Feutrill, Hilary Boucaut, Francois Duvenage, Tim Boucaut, Peter Hewett, Sanjeev Khurana Health Data Analytics 2019, Sydney, October 16-17 2019
Disclosures www.datadissect.com.au Health technology company Collaborative of Surgeons, Mathematician, IT Consultant, Computer Scientists, Business Manager
Limitations with current EBM • ‘Status quo’ – Linear process – Start and end dates of study/trial – Strict inclusion/exclusion criteria – Generalisability of results not guaranteed – Expensive – Inefficient – Translation to practice not guaranteed • Health Knowledge creation is an industry in itself and not by-product of clinical care • Disconnect between administrators, researchers and clinical work force
Medical research landscape is changing • Scientific method based on reductionist science – Isolating outcome measures or variables to investigate causes or effects • “Research is changing from a hunter/gatherer mode, where huge amounts of effort is invested to associate data with rare events, to a harvest mode in which huge amounts of data are used more efficiently to give insight.” – Embrace multi-dimensional, multi-disciplinary data with human-computing symbiosis 1. http://www.learninghealthcareproject.org/publication/5/66/dr-paul-wallace-interview
Value-based healthcare Value = (Quality + Outcomes + Safety) Cost
Variation in care • E.g. surgical management of urinary reflux in children Ureteric reimplantation rate (per 100,000) NSW 1.0 VIC 0.8 QLD 2.5 SA 6.5 WA 10.8 TAS 0
Learning Healthcare System • Facilitated by the Electronic Medical Record paradigm – 2007 first description – 2013 Institute of Medicine definition • “Any type of healthcare delivery system that combines research, data science, and quality improvement, yielding knowledge as a by-product of the patient-clinician interaction and focused on improving patient health and system outcomes” • Health sector slow to adopt from concept to action – Immense and rapidly changing volume of medical information, complexity of decision making, limited capacity to evaluate decisions • Only 13 publications reporting actual implementation (2016) 1. Deans KJ, et al. Learning health systems. Semin Pediatr Surg. 2018 Dec;27(6):375-378. 2. Forrest CB, et al. Development of the Learning Health System Researcher Core Competencies. Health Serv Res. 2018 Aug;53(4):2615-2632. 3. Budrionis A , et al The Learning Healthcare System: Where are we now? A systematic review. J Biomed Inform. 2016 Dec;64:87-92. 4. Kwon S, et al. Creating a learning healthcare system in surgery: Washington State's Surgical Care and Outcomes Assessment Program (SCOAP) at 5 years. Surgery. 2012 Feb;151(2):146-52.
What is a Learning Healthcare System? • Continuous improvements in quality, outcomes & efficiency – Cycle begins and ends with clinician-patient interaction – Improving rather than proving – Afferent (blue) and Efferent (red) arms – Research influences practice, and practice influences research • Distinguishing features – Patient/family engagement through self-reported outcomes – LHS researchers embedded at point-of-care – Leverages evidence about “what works” in context of own setting 1. Deans KJ, et al. Learning health systems. Semin Pediatr Surg. 2018 Dec;27(6):375-378. 2. Greene SM, et al. Implementing the learning health system: from concept to action. Ann Intern Med. 2012 Aug 7;157(3):207-10. 3. http://www.learninghealthcareproject.org
Differences between LHS & Clinical Registry/Audit Clinical information Learning Health System Medical Record Registry or Audit • Equally stringent data acquisition and storage • Less amenable to point of care data entry • Multi-user cloud based platform • Expensive • All data formats including images and videos • Often requiring salaried data entry staff • Timely insight into outcome and process of care • Lag phase in outcome reporting • Single platform to store patient info leaflets and • Limited potential for actionable insight for consent, etc quality improvement
Fundamental building blocks Clinicians Tech platform Data scientists Context Acquisition Interpretation Data Collation Visualisation Storage Analytics Socio-technological system dependent on technical underpinnings
Fundamental building blocks Socio-Technological System Human factors Technological factors • • Motivated stakeholders with desire to System-wide accessibility allowing continuously improve system learning to permeate organization • Willingness to be vulnerable and transparent • Clinical leadership • Domain experts • Administrative support • Ability for patient to actively participate in their data collection via QoL assessments that populate platform
Washington example • SCOAP (Surgical Care & Outcomes Assessment Program) – Launched 2006 – 60 out of 65 hospitals with surgical service – Surgeon designed, grassroots, voluntary, peer-based QI collaborative • Creates value proposition for surgeons and hospitals to join • Initially appendicitis, colorectal surgery, bariatric surgery – SCOAP OR Checklist – Decreased negative appendicectomy – Decreased UTI in epidural patients – Decreased anastomotic leak rate in colorectal surgery – Decreased blood transfusions (only if Hb < 7 g/dL) – Improved nutritional pre-habilitation for elective surgery – Appropriate use of neoadjuvant therapy for rectal cancer 1. Kwon S, et al. Creating a learning healthcare system in surgery: Washington State's Surgical Care and Outcomes Assessment Program (SCOAP) at 5 years. Surgery. 2012 Feb;151(2):146-52.
Introducing LHS ‘culture’ for quality improvement 2014 1. Health Roundtable national audit data. https://www.healthroundtable.org/
Introducing LHS ‘culture’ for quality improvement • Paper based data collection and machine learning – Fast track protocols – Identification of a subset of patients with complicated appendicitis that could be safely discharged earlier Appendicitis “Uncomplicated” “Advanced” “Complicated” 15
Introducing LHS ‘culture’ for quality improvement 2014 2017 $143,803 saving p/a 1. Health Roundtable national audit data. https://www.healthroundtable.org/ 2. Cundy TP, et al. Fast-track surgery for uncomplicated appendicitis in children: a matched case-control study. ANZ Journal of Surgery. 2017 Apr;87(4):271-276.
Paper-based v1 Digital LHS platform • Bespoke LHS digital platform – Customised user interface – Browser and smart-phone functionality – Hosted on off-site University of Adelaide server – Approval for satisfying patient privacy and data security protocols
Clinical implementation • Integrated into day-to-day practice – n = 180 consecutive patients in 9 months – Multi disciplinary teams gather data that is not currently being captured at the point of care • Structured and unstructured data (including media)
Future work • Scale-up • Patient reported outcomes • Data visualization • Growing demand for applications
Conclusions 1. Established first active LHS in Australia 2. Continual flow of actionable cloud-based data for analysis – Research influences practice, and practice influences research 3. Healthcare services stand to benefit – Rapid access quality improvement – Improved patient outcomes with reduced cost of care
Thank you www.datadissect.com.au Health technology company Collaborative of Surgeons, Mathematician, IT Consultant, Computer Scientists, Business Manager
Recommend
More recommend
