AI Use Cases 

Provide support for medical teams to triage incoming patients relative to their medical symptoms and history and personal situation. Data used may include audio, video, comparable, historic and real-time data.

Predict personalised health outcomes to optimise an individual's recommended treatment package. The data required to support this is often still being understood and developed.

Analyse body or breath odour to diagnose potential disease. Sensitive tools can be used to capture and then analyse odours which are beyond the range of the human olefactory system.

Re-examine data from historic research to discover new applications. Traditional methods may not have captured all the complexity that AI can parse - or the data may indicate that new techniques and technology would be applicable to the data.

Gene analytics and editing can be sped up and delivered with high accuracy using AI. The potential for radical change in patient outcomes in oncology is one of the most interesting potential outcomes from applied AI.

Keeping track of various treatment options - prescriptions, plans etc using sensors to automate billing and administration

Medication (dosage) effectiveness will potentially vary by individual and circumstance. Machine learning can help build individualised healthcare pathways.

Predict likely outcome risk for patient to support healthcare providers and affected individuals (and family) with planning, correct treatment pathways and recommended responses, including palliative care support.

Identify fraud, waste, and abuse patterns from a variety of clinical operations data. This would typically occur at a cross-functional level, but could be at an individual, operating unit or institutional level.

Predicting changing health patterns in a population based on data including which health services are accessed and how often, patient treatment outcomes, demographics etc. Allows for healthcare system resource planning.

Predict prescription addiction and abuse using data from both individuals and broader historic pattern data. This is especially impactful where there are clusters of cases in similar locations.

Automated medical response to individuals with health questions available through media including apps, online and chat functionality. Potentially used as part of a triage functionality in a healthcare system.

Behavioural nudges (eg reminders) to encourage prescription intake and positive steps towards healthy habits specifically for managing chronic health conditions. Personalisation of messaging pathways will help those in what may be a painful and lonely situation (although not as much as sympathetic personal contact....).

Assist physicians by providing the latest medical information and research to inform proper patient care - this may be from specialist providers and cover the sorts of rare ailments that medical staff may not encounter on a regular basis

Detect major trauma events from wearable sensor data and signal emergency response. Real time sensors are used to monitor individual symptoms to alert emergency staff to urgent health issues. This can be extended to analysing incoming phone calls for audible clues.

Claims review prioritisation to ensure that cases are handled effectively and efficiently.

Diagnose injury (eg brain trauma) from visual scans

Personalise messaging and approach to improve wellness and adherence to prescriptions and medical programmes - this can be delivered as a 'nudge' programme

Monitor and react to health symptoms in post-hospitalisation care programmes, using wearable sensors and data captured on smart devices. Machine analysis of the data will potentially trigger alerts to medical personnel for active intervention.

Consumer-ported real time sensor devices (phones, watches etc) provide data for remote analysis. Analysis provided can be personal or via medical practitioner.

Monitor data from biopsys to forecast likely medical implications. Rate of turnaround and accuracy of predictions are key factors.

Medical treatment plan optimised for an individual based on an analysis of their genetic data. The science is still at an early stage but deployment of AI will help speed up development.

Scan and replicate human body parts for more accurate implants. This technology has potential to work with teeth, bones and other implants that need precision production to minimise any negative impact on the host body. There may be production benefits achievable in the future.

Using NLP to describe the surrounding data collected and compiled using smartphones and Deep learning to help people with varied degrees of blindness navigate

Defining tumour contours versus the surrouding healthy tissue to enable accurate and limited radioactivity deplpoyment is currently a manual task. This is painstaking work that is time-consuming and (given the cost of highly skilled surgeons doing the work) expensive which means that is rarely gets done between individual treatments and doctors approaches can differ (raising treatment risks). AI can deliver this at speed and economcially - which speeds up the treatment process and can be repeated.

Initially screening health insurance claims to detect incorrect claims

Sorting insects by sex enables aggressive intervention (e.g. sterilisation or culling) to minimise their ability to procreate and therefore spread disease. Machine vision enables this to happen at greater scale and pace therefore enabling more significant interventions. This also has research benefits.

Optimise pricing strategy for drug portfolio. Whilst this will likely increase portfolio yield there are reputational risks if mis-managed.

Alerting and diagnostics from real-time patient data - this might include, for example, warning of a life-threatening incident such as a heart attack.

Optimise resource allocation in drug development using both internal and external (e.g. social media) data.