{{item.videoDuration}}
{{item.title}}
{{item.videoDuration}}
The computer will see you now: Six examples of AI in healthcare
Key takeaways
- Artificial intelligence can improve efficiency of hospital administration, allowing doctors to focus on patient treatment.
- Transforming health data for analysis will help improve diagnosis and treatment.
- Virtual services can offer a significant reduction in financial costs for healthcare providers.
As an industry defined by the relationship between patient and carer, at first glance it may seem incongruous to nudge healthcare towards a robotic future. In fact, artificial intelligence (AI) has the potential to completely reshape the health industry, offering greater support to human capabilities and allowing healthcare organisations to deliver higher-quality services more efficiently.
Artificial intelligence is a broad term for computer systems that can ‘think’ and act like humans. They can sense their environment, absorb information, learn from past experience, make decisions and take action.
AI has transformative power for two reasons: the explosive growth in data, coupled with huge computational advances and processing speeds. In healthcare, the potential for disruption is limitless. It is an industry that generates phenomenal amounts of data, such as patient case histories, population-level datasets, real-time digital data from smartphones and medical wearable devices, as well as unstructured sources such as CT scans and doctors’ handwritten notes.
Technology transforming the industry
Implementing AI as part of a data-driven approach to healthcare offers three big opportunities. The first is achieving new excellence in service delivery. The second is meeting rising citizen expectations for personalised healthcare experiences. Third, governments can design a more efficient system, lowering the per capita cost of managing a rapidly growing and ageing population.
Many healthcare organisations around the world have already integrated artificial intelligence into their services and processes, from back-end management of hospital administration, through to patient diagnosis and treatment. Here, we look at some examples of AI in action:
Keeping track of supplies
An Australian health provider is working with PwC to improve the efficiency and service levels of its medical consumables supply chain. Over 7,000 wards were categorised based on their individual weekly purchasing behaviours. Forecasting showed that over 70% of these orders could be automated with seven days’ worth of safety stock. This shift to automated ordering will improve accuracy and lead to better demand management.
Turning health records into real-time responses
The reporting of cancer cases to cancer registries remains a paper-based process here in Australia, which means that statistics about cancer incidence are often outdated. In combination with Queensland Health, the Australian e-Health Research Centre is working to transform free-text health data from pathology reports into computable, ready-to-use structured data for AI applications. The aim is to establish a real-time cancer registry that can provide cancer incidence data to inform activities such as monitoring, health service planning and research.
Forecasting hospital demand…
The Queensland Government is using the CSIRO’s Patient Admission Prediction Tool (PAPT) to forecast the demand on hospital beds, staff resources, and elective surgery in a bid to cut patient waiting times at 27 major hospitals across the state. The software analyses historical data to predict, with around 90% accuracy, how many patients will present at emergency departments and when.1 It also predicts a patient’s medical needs and urgency of care. PAPT is now being extended to predict diseases such as influenza and the hospital admissions of patients with chronic diseases.
… and reducing readmissions
A rural health district in NSW was experiencing high rates of unplanned hospital readmissions, which require enormous financial and staffing resources to manage. The hospital sought an AI solution to detect which patients were at a high risk of readmission. It integrated more than a decade of medical records to create an AI-powered machine learning model that considered patient characteristics such as demographics, case histories, and treatment results. By doing so, it was able to identify, with 70% accuracy, patients who had an unexpected readmission within 28 days.2 This information is now being used to help doctors make real-time decisions that improve care quality.
Monitoring patients with ‘virtual nurses’
Leading healthcare providers in the UK have trialled an AI-powered app to monitor post-operative patients and those living with chronic illnesses. Virtual nurse ‘Molly’ checks in with patients, answers their health questions and pulls data from a range of devices that can be accessed by healthcare professionals. The app can also detect changes in patients’ moods and identify when they are experiencing feelings of depression or anxiety as a side effect of medication or lifestyle changes. It is estimated that virtual health assistants could save US$20 billion annually by reducing communication between nurses and patients by 20%.3
Improving treatment for patients
Dementia Support Australia is using AI to identify the presence and severity of pain in patients with dementia who are no longer able to express their discomfort. The tool, known as PainChek, runs analysis on a 10-second video of an individual’s face, with the aim of detecting pain-related expressions including brow lowering, cheek raising, tightening of the eyelids and wrinkling of the nose.
This is an exciting but daunting time for healthcare providers. While adopting AI may seem complicated, case studies such as those outlined above prove that the technology has limitless potential to disrupt the industry.
Everyone, from doctors to nurses and support staff, needs to accept that physical healthcare environments will change, and they must embrace new skills and ways of working. At the same time, healthcare will always require a human touch. Those incorporating AI should never lose sight of the fact that technology best supports human capabilities, rather than replaces them, and that compassion, intuition and emotional intelligence will remain pillars of true healthcare.
For more detailed information on AI and its implementation in the healthcare industry, read Adopting AI in healthcare: Why change?
References
- https://www.csiro.au/en/Research/BF/Areas/Digital-health/Waiting-times
- https://aibrisbane.com.au/casestudy/using-machine-learning-to-predict-hospital-readmission/
- https://hbr.org/2018/05/10-promising-ai-applications-in-health-care
{{filterContent.facetedTitle}}
{{contentList.loadingText}}
{{contentList.loadingText}}